Classroom Companion: Business

The Classroom Companion series in Business features foundational and

introductory books aimed at students to learn the core concepts,
fundamental methods, theories and tools of the subject. The books offer
a firm foundation for students preparing to move towards advanced
learning. Each book follows a clear didactic structure and presents easy
adoption opportunities for lecturers.
More information about this series at http://​www.​springer.​com/​
series/​16374
Harald Øverby and Jan Arild Audestad

Introduction to Digital Economics

Foundations, Business Models and Case
Studies
2nd ed. 2021
Harald Øverby
Norwegian University of Science and Technology, Gjøvik, Norway

Jan Arild Audestad
Norwegian University of Science and Technology, Gjøvik, Norway

ISSN 2662-2866 e-ISSN 2662-2874

Classroom Companion: Business
ISBN 978-3-030-78236-8 e-ISBN 978-3-030-78237-5
https://doi.org/10.1007/978-3-030-78237-5

Originally published by Create Space Independent, Scotts Valley, CA,

USA, in 2018 with ISBN 978-1986751391

© The Editor(s) (if applicable) and The Author(s), under exclusive

license to Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are solely and exclusively
licensed by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in
any other physical way, and transmission or information storage and
retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks,

service marks, etc. in this publication does not imply, even in the
absence of a specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general
use.

The publisher, the authors and the editors are safe to assume that the
advice and information in this book are believed to be true and accurate
at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the
material contained herein or for any errors or omissions that may have
been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer

Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham,
Switzerland
To Janny and Synnøve
Foreword
It is my great pleasure to introduce this important book. While the
impacts of digital technologies are growing with far-reaching
magnitudes, the academic field of digital economics is still in its
inception and definition phases, with even fewer textbooks. This book
fills an important void by bringing together theories from economics,
management science, system dynamics, and business modelling to
provide overview and structure to this emerging field. The book will be
useful both for graduate and under-graduate courses, and for
researchers who wish to have a salient overview and introduction.
The book has a clear pedagogical structure and flow, where each
chapter features abstract, learning objectives, definitions, logical
partition with sub-headings, theory presentations, cases, boxes, and
concluding set of Q&A. With this lay-out, the book is immediately
appealing for parsing and selection of chapters that seem more
interesting. The order between chapters is also clear, one building upon
each other, starting with the technologies, and moving step by step into
more and more sophisticated topics. Toward the end, there are chapters
with quite advanced mathematical modelling, together with some
specific application areas such as big data and network neutrality. The
final chapter provides an overview of digital regulation, as a final
integrating chapter that brings to bear many of the economic analysis
tools presented earlier.
While organizing and summarizing theories of relevance for the
topic, the book also has some important intellectual contributions and
perspectives. Among others, the book emphasizes the integration of
Porter’s strategy framework with the value network perspective which
serves to integrate both traditional strategy analysis with new theories
relating to value creation in digital networks. Moreover, the integration
of the multi-sided platform theories with the value network theory and
network effects is an important way forward to operationalize the
multi-sided platform theory further. The authors have developed a set
of excellent illustrations to further illustrate the importance of the
various relations between multi-sided platforms with other theories
and constructs. The book also provides a helpful historical context by
referring the readers to origins and authors of important concepts,
such as ecosystems and path dependencies, and by so doing providing
teasers for further reading.
In summary, the book provides an integrated overview of a field of
increasing importance and has done so in a very pedagogical fashion.
The book should be a welcome contribution for courses in strategy,
business, and digital economics. The book is suitable for both
undergraduate and graduate level engineering students, taking
advantage of the advanced mathematical modelling.
I give this book my heartfelt recommendation and invite the reader
to enjoy this tour of digital economics, in the company of two experts
and enthusiastic guides!
Erik Bohlin
April 14, 2021
Preface
This book is an introduction to digital economics. It is highly cross-
disciplinary and draws upon knowledge from several academic
disciplines such as telecommunications, computer science,
management science, business modeling, economics, and mathematics
to explain the digital economy. To fully comprehend digital economics,
it is important to understand how the information and communication
technology (ICT) is underpinning all digital businesses.
The six biggest companies by capitalization (by 2021) are all major
stakeholders in the digital economy. The digital economy can no longer
be ignored as an oddity in economy theory. On the contrary, the
economics of digital goods and services has become a key element in
the understanding of how the world economy works.
Several topics covered by this book are included in the curricula
taught at several business and economics schools around the world.
This book is unique as it approaches topics in digital economics from a
technology point of view. We believe that this is essential since digital
economics is a result of the evolution of information and
communications technologies, and not vice versa.
The field of digital economics is complex and cannot be fully
understood using theories from traditional micro-economics alone. It is
necessary to adopt existing theories using knowledge from system
dynamics, management science, and business modelling.
We wrote this book to support the growing community of students
and practitioners with textbook material that links the theoretical
foundations of digital economics with practical examples and case
studies.
Harald Øverby
Jan A. Audestad
Gjøvik, Norway
Gjøvik, Norway
May 2021
About this Book
This book is about digital economics—the branch of economics
studying digital goods and services. Innovations and developments in
information and communication technology (ICT) have laid the
foundations for this branch of the economy. This includes technologies
such as social media, apps, cloud computing, mass storage, data mining,
cryptocurrencies, and sharing services. These services have already
made deep imprints on today’s business landscape. Both private
businesses and the public sector embrace ICT to achieve cost benefits,
efficiency, and competitive advantages. However, what we are
witnessing is just the beginning of an economic revolution as the full
potential of the digital economy is about to be harvested.
The major goal of this book is to provide a theoretical basis for
digital economics and to show how these theories can be applied to the
study of real-world economics and business phenomena. The book is
cross-disciplinary, explaining how the interaction between markets and
important innovations in telecommunications and information
technology has shaped the digital economy. The field of digital
economics is characterized by transient market behavior, feedback
mechanisms, global markets, many stakeholders, and technology
dependencies never seen in any markets before. The book highlights
this complex ecosystem.
The book is written for advanced undergraduate courses in digital
economy for students in computer science, economics, and
management. To get the full benefits from the book, a student needs
first courses in computer science, calculus, and economics. The book
may also be of interest to a broad range of professionals, including
economists, business consultants, managers, and computer scientists
since it studies the impact several technological, social, and economic
disciplines have on the evolution of the digital economy.
Each chapter may contain, in addition to the main body of the
chapter, elements such as:
– Learning goals are at the beginning of the chapter.
– Case studies are real-world examples of digital products and markets.
– Examples are used to show concrete applications of the theoretical
material not associated with a particular real-world case.
– Boxes contain supplementary material. In particular, mathematical
derivations are contained in boxes to make the text more accessible.
– Questions and answers are included at the end of each chapter. The
questions are not direct repetition of elements in the text but may
require search for information in other sources, e.g., the web.
– Further reading. Some chapters contain references to books or
articles providing deeper insight into the material presented in that
chapter.
The book consists of 22 chapters. These chapters can be grouped
into several themes:
– ► Chapter 1 is an introduction where the basic concepts of the
digital economy is defined.
– The technological evolution of digital networks, goods, and services is
the topic of ► Chaps. 2, 3, and 20. ► Chapter 2 shows the timeline of
major inventions in the information and communication technology
leading to the digital marketplace we see today. ► Chapter 3 shows
how the communication network converges into a single network
based on the Internet technology and, thus, replacing dedicated
networks such as the telephone network. ► Chapter 20 contains
supplementary material on the impact of the big data technology on
the digital economy – big data is a core technology for several of the
biggest companies in the digital economy. Technology is a side issue
in almost all chapters since almost all aspects of the digital economy
are intimately connected to the technology.
– Key characteristics of the market are identified in ► Chaps. 4, 5, 6, 7
and 18. ► Chapter 4 describes the ecosystem of digital markets,
including the various stakeholders that shape the business
landscape. This chapter also shows that the protocol structure of the
Internet divides the digital businesses into two separate business
categories: Internet service providers and application and content
providers. This division is the key element behind the enormous
innovation of digital applications since 2000. ► Chapter 6 is one of
the most important chapters in the book since it defines the basic
characteristics of digital goods and services and what makes them
 different from physical goods. The most important characteristic is
that the marginal cost of digital products is zero, allowing companies
to offer digital services free of charge and thereby generating a
tremendous market for them. Combined with big data technology,
these companies may then exploit the data they collect about their
users to generate revenues from advertisements and other products.
► Chapter 5 describes how the telecommunications market evolved
from de facto monopolies to full competition, and ► Chap. 7 outlines
new production models for digital services enabled by the
technological evolution. ► Chapter 18 contains supplementary
mathematical models describing how these markets evolve as a
function of time.
– Various strategic issues are contained in ► Chaps. 8, 9, 10, 11, 12, 13,
14, 15, and 16. ► Chapter 8 discusses how value is created in
different businesses: chains producing physical goods, shops solving
problems for clients, and networks acting as mediators between
groups of clients. One particularly important strategic element is
network effects, or positive feedback from the market. This is
explained in ► Chap. 9. ► Chapter 10 about multisided platforms is
an application of the theories of ► Chaps. 8 and 9. The other
chapters in this groups are concerned with other strategic issues
such as path dependence (11), lock-in (12), formation of monopolies
and oligopolies (13), acquisitions and mergers (14), development of
technical standards (15), and the long tail (16).
– ► Chapters 17 and 19 are about business modeling. ► Chapter 17
characterizes the digital markets as e-commerce markets, network
access markets, and information services markets. ► Chapter 19
demonstrates, using several real-life examples, the use of the
business model canvas and the stakeholder relationship model as
powerful tools for business planning and analysis.
– Regulations are the subject matter of ► Chaps. 21 and 22. ► Chapter
21 is about the complex political issue of net neutrality. Net
neutrality implies non-discriminate access to and use of the Internet.
In some countries, this is an obligatory regulation while not in other
countries. ► Chapter 22 is about other types of regulations
concerning provision and use of the communications infrastructure.
Abbreviations
3G Third Generation Mobile System
4G Fourth Generation Mobile System
5G Fifth Generation Mobile Systems
ADSL Asymmetric Digital Subscriber Line
ARPU Average Revenue per User
ASP Application Service Provider
B2B Business to Business
B2C Business to Consumer
BMC Business Model Canvas
BPQ Buyer-Player-Quitter
C2B Consumer to Business
C2C Consumer to Consumer
CBPP Commons-Based Peer Production
CP Content Provider
CPU Central Processing Unit
DAB Digital Audio Broadcasting
EDGE Enhanced Data rates for GSM Evolution
EEA European Economic Area
FinTech Financial Technologies
FTTH Fiber to the Home
GPRS General Packet Radio Service
GSM Global System for Mobile Communications
HTML Hypertext Markup Language
HTTP Hypertext Transfer Protocol
IaaS Infrastructure-as-a-Service
ICT Information and Communication Technology
IoT Internet of Things
IP Internet Protocol
ISP Internet Service Provider
ITS Intelligent Transport Systems
ITU International Telecommunication Union
LTE Long Term Evolution
M&A Mergers & Acquisitions
MC Marginal Cost
MMOG Massive Multiplayer Online Game
MMS Multimedia Messaging Service
MOOC Massive Open Online Course
MSP Multi-sided Platform
MVNO Mobile Virtual Network Operator
NFV Network Function Virtualization
NMT Nordic Mobile Telephone
NP Network Provider
O-T Odlyzko-Tilly
OTT Over-the-Top Services
PaaS Platform-as-a-Service
PC Personal Computer
PDF Portable Document Format
PSTN Public Switched Telephone Network
SaaS Software-as-a-Service
SCTP Stream Control Transmission Protocol
SDN Software Defined Networking
SIR Susceptible-Infectious-Recovered
SLA Service Level Agreement
SMS Short Message Service
SOA Service Oriented Architecture
SRM Stakeholder Relationship Model
TCP Transmission Control Protocol
UDP User Datagram Protocol
URL Uniform Resource Locator
VCR Videocassette Recorder
VHS Video Home System
VNO Virtual Network Operator
VoIP Voice over IP
VoLTE Voice over LTE
WoW World of Warcraft
WWW Word Wide Web
XaaS Anything as a Service
Contents
1 The Digital Economy
2 Information and Communication Technologies
3 Convergence of Technologies and Services
4 Digital Economy Ecosystem
5 Digital Market Evolution
6 Digital Goods and Services
7 Production Models
8 Value Creation Models and Competitive Strategy
9 Network Effects
10 Multisided Platforms
11 Path Dependence
12 Lock-In and Switching Costs
13 Digital Monopolies and Oligopolies
14 Mergers and Acquisitions
15 Standards
16 The Long Tail
17 Digital Markets
18 Digital Market Modeling
19 Digital Business Models
20 Big Data Economics
21 Net Neutrality
22 Digital Regulation
Index
About the Authors
Harald Øverby
(born 1979) is a professor at the Norwegian University of Science and
Technology (NTNU). He received his M.Sc. in computer science in 2002,
a B.Sc. in economics in 2003, and a Ph.D. in information and
communication technology in 2005. Øverby has also studied law at the
University of Oslo in Norway. Øverby has published over 90 papers in
international journals and conferences in the areas of communication
technology, digital economics, business modeling, data coding, and
optical networking. He has taught and developed university courses in
the areas of digital economics, business, and law since 2010. He was the
Head of Department of Telematics (NTNU) from 2013 through 2016.
Øverby is a 1998 International Chemistry Olympiad bronze medalist.

 
Jan A. Audestad
(born 1942) is professor emeritus at the Norwegian University of
Science and Technology (NTNU). He received his M. Sc. in theoretical
physics in 1965. He has more than 50 years of experience in
telecommunications both from academia and as researcher, research
manager, and techno-economic adviser to the top management of
Telenor. As adjunct professor of NTNU, he has taught courses in
telecommunications networks, distributed processing, mobile
communication, digital economics, and information security. He was
one of the key scientists behind standardization of maritime satellite
systems in the 1970s, GSM mobile communications in the 1980s, and
intelligent networks and distributed processing in the 1990s. He was
part of the team reorganizing Telenor to meet full competition in all
sectors of telecommunications from 1995 to 2009.

 
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_1

1. The Digital Economy

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Digital economy – Digital economics – Digitization – ICT

Learning Objectives
After completing this chapter, you should be able to:
– Understand the size and versatility of the digital economy.
– Explain how the adoption of Internet access and mobile
telephony has enabled the digital economy.
– Understand what is meant by digitization of the economy and
how it is related to digitization of communication networks
and production and storage of digital information.

1.1 Introduction
Information and communication technology (ICT) is everywhere
around us—the Internet, smartphones, laptops, wireless networks,
apps, and online video services such as Netflix and HBO. ICT has
become ubiquitous, at least in the developed world. The pace of
innovation in ICT is fast, and new technologies are emerging every year.
Over the last few decades, ICT has changed how we work, how we
spend and invest our money, and how we conduct our business.
Telecommunications, finance, and media are industries in which ICT
has significantly changed the business landscape.
Spotify and other providers of online music services have radically
changed the business models of the music industry, particularly,
reducing revenue streams from CD retail. Since an increasing amount of
music is traded online, the need for physical stores selling CDs has
almost vanished. Worldwide sales of recorded music have decreased by
45% from 1999 to 2014. The year 2014 also marked the first year in
which online traded music matched sales from physical formats such as
CDs (Reid, 2015).
E-banking has radically changed how we—as consumers—
approach banks and other financial institutions. Most activities
involving personal finance are now conducted over the Internet using
smartphones or personal computers. For active e-banking users, there
is no need to visit a physical bank to pay bills. Loans can be negotiated
with the bank over the Internet. Cash is no longer needed to pay for bus
or train tickets, car parking, or taxis.
At many airports, passengers check in automatically, put their
baggage on the baggage drop belt, and pass through the gate to the
airplane without the involvement of ground personnel. All passenger
services are completely automatic, except for the security control.
It is the increased use of digital goods and services, often as a
replacement for physical goods and non-digital services, which is
responsible for this evolution. In fact, digital goods and services are the
essential building blocks of the digital economy. Even though digital
goods and services have gradually changed the business world for some
time already, we are just now seeing the beginning of an economic
revolution as the full potential of the digital economy is about to be
harvested.

1.2 Definitions
Definition 1.1
The digital economy
is an economy based on information and
communication technology (ICT).
The digital economy is based on information and communication
technologies such as the Internet, smartphones, mobile and wireless
networks, optical networks, Internet of Things (IoT), cloud storage and
cloud computing, sharing services, apps, and cryptocurrencies. The size
and impact of the digital economy are driven by people’s adoption of
these technologies.
◘ Figure 1.1 shows the share of households with access to the
Internet for the period 2005–2019 (ITU statistics, 2018). In 2005, only
20% had access to the Internet. Fourteen years later, in 2019, about
60% of the world’s population has access to the Internet. Over the last
decade, access to the Internet has increased in all parts of the world.
However, there are significant differences—a digital divide—in Internet
adoption between countries and within countries. While about 85% of
households in the developed has access to the Internet, less than 50%
of the households in the developing world has access to the Internet. A
key question for the next decades is how to provide Internet access to
the developing parts of the world.
Fig. 1.1 Percentage of households with the Internet for the period 2005–2019. (Authors’ own
figure)

An important evolution in the digital economy is the number of

people using public narrowband and broadband mobile technologies.
Cellular narrowband mobile systems (2G
) offer global services such as
telephony and SMS. Cellular broadband mobile systems (4G and 5G)
support the use of smartphones to access the Internet. These
technologies also support telephony and SMS, eventually phasing out
the use of 2G and 3G systems. The number of users of public mobile
networks has surpassed the number of people in the world—in 2020
there are 105 active mobile cellular subscriptions per 100 inhabitants
in the world. The reasons for this are that many people have access to
more than one device (e.g., one private smartphone and one for work)
and that mobile phones are used as autonomous communication
devices for connecting sensors and other devices on the Internet of
Things (IoT) and the public infrastructures.
◘ Figure 1.2 shows the access to 4G/LTE mobile networks for the
period 2015–2020. Observe that even in the least developed countries,
about 40% of the population has access to 4G/LTE networks. For the
same group of countries—i.e., the least developed countries—close to
90% of the population has access to mobile cellular networks, and
about 75% of the populations has access to 3G networks. Hence, access
to the mobile telephony is more widespread than access to the Internet.

Fig. 1.2 Worldwide access 4G/LTE mobile networks. (Authors’ own figure)

Access to the Internet has been proposed as a basic human right,

and in 2016, the United Nations (UN) released a non-binding resolution
condemning intentional disruption to such access by governments
(Vincent, 2016). It is clear that access to the Internet has changed the
lives of people and the way businesses operate and will increasingly do
so as the other half of the world’s population gets connected to the
Internet.

Facets of the Digital Economy

Three representative examples of digital goods and services are
presented to indicate the extent of the digital economy. Many more
examples are provided throughout this book.
Facebook
: There are over three billion people worldwide using a
social media service regularly. Facebook is the most popular of these
services, with approximately 2.2 billion users. Facebook has had a
significant impact on how people communicate and organize their
social lives. The advertising industry has been dramatically changed
because of the way social media advertisements can be tailored to
match the attitudes and preferences of the users. Because of the
company’s global impact, Facebook founder, Mark Zuckerberg, was
named “Person of the Year” by Time Magazine in 2010. However,
Facebook is not without controversies, as witnessed by the
Cambridge Analytica data scandal in 2018.
Airbnb: Established in 2008, Airbnb has grown to become one of
the biggest hospitality services worldwide. As of 2017, Airbnb has
over 200 million users and offers over three million lodgings in 191
countries. Airbnb enables homeowners to rent out property to
registered guests and is one of the best examples of the expanding
sharing economy. Airbnb utilizes the concept of multisided platforms
(► Chap. 10) and the long tail (► Chap. 16) in its business
operations. Key enablers for Airbnb’s success are the widespread
adoption of Internet access, high-speed mobile networks, and
smartphones. However, Airbnb has been blamed for reducing
attractiveness in the neighborhoods in which it operates and has met
resistance from authorities in, for example, Paris and New York.
Bitcoin: Bitcoin
(BTC) was established in 2009 by the still-
unknown person or organization Satoshi Nakamoto. It has become
the most valuable and well-known cryptocurrency. Bitcoin utilizes the
blockchain technology to provide a distributed currency without any
involved third parties. Bitcoin has the potential to become a true
 global currency. However, recent investigations have revealed several
weaknesses with Bitcoin, including long transaction times and high
energy usage. Other cryptocurrencies, such as Litecoin
(LTC),
Ethereum
(ETH), and Ripple
(XRP), may overcome these challenges
and turn out to be the dominant cryptocurrencies of the future—if
cryptocurrencies have any future at all.

The size of the digital economy is hard to estimate. This is because ICT
is an industry on its own (production of telecommunications networks,
Internet equipment, mobile phones, applications, and software), but
also because ICT is integrated into almost all other industries. ICT has
enabled new business models, more efficient production methods, and
new ways of interacting with consumers. One example is online trading
(e-commerce) where people can buy almost any kind of merchandise
using the Internet. We are experiencing a transition from the industrial
economy to the digital economy—from physical products to digital
goods and services.
The world’s six largest corporations by market capitalization as of
the first quarter of 2021 (in descending order) are Apple, Microsoft,
Amazon, Alphabet (Google), Facebook, and Tencent (Wikipedia, 2021)
.
All of these companies produce digital goods and services and are
major businesses in the digital economy. As of March 30, 2021, their
combined market capitalization totals more than $8300 billion. ◘
Figure 1.3 illustrates the market cap of the top ten companies in the
world.
Fig. 1.3 Top ten corporations worldwide according to market cap. (Authors’ own figure)

These companies hold immense power in today’s business world

due to their size, span of operations, and international impact. They can
be characterized as digital conglomerates as their business operations
have expanded far beyond their original business idea (Tapscot et al.,
2006). Google, for example, started out as a company delivering search
engine services for Internet users. Today, Google offers, in addition to
its search engine, e-mail (Gmail), instant messaging, learning platform
(G Suit for Education), voice-over-IP (Google Hangouts), text editing
(Google Docs), cloud storage (Google Drive), and several other
platforms. Google has expanded its business operations into many
sectors of the digital economy by acquiring competing companies and
by horizontal and vertical integration.
The main asset of these companies is the network of consumers of
the digital goods and services they offer. These users give rise to
network effects
(see ► Chap. 9) that generate huge value to these
companies. Maybe the most striking fact is that these companies have
only needed 20 years to gain their current market dominance. In 2008,
the top five companies according to market capitalization included
PetroChina, ExxonMobil, General Electric, China Mobile, and Industrial
and Commercial Bank of China (ICBC). Of these companies, only China
Mobile can be said to fully operate in the digital economy by offering
Internet and mobile access to consumers in China. As we saw above, the
six top businesses in 2021 all belong to the digital economy, and one of
them (Tencent) has become the biggest company in China. How do
companies in the digital economy get so big? How is it possible for
companies in the digital economy to accumulate so much value in such
a brief time? These are some of the questions we will shed light on in
this book.
Several disruptive innovations
have contributed to the scope and
size of the digital economy. A disruptive innovation is an innovation
that creates a new market, often leading to a change in market
leadership and the emergence of new companies which become the
new market leaders (Christensen, 1997).

Failure to Accept Disruptive Technology

One of the most famous examples of disruptive innovations is the
failure of the photography company Kodak. Kodak was one of the
leading producers of chemical photography and camera films. Kodak
failed to embrace digital photography in the 1990s and 2000s and lost
the competition with Asian producers. Kodak filed for bankruptcy in
2012, and its patents were bought by a group of companies (including
Google and Apple) for $525 million in 2013. Later that year, Kodak
emerged from bankruptcy; however, with a vastly different market
position compared to the market leader it once was in the 1990s.

When ICT is at the core of a disruptive innovation, the market often

changes from producing physical products to producing digital goods
and services. Market sectors that have been significantly affected by
ICT-based disruptive innovations are media, telecommunications, and
finance (Grossman, 2016).
An important part of the digital economy is e-commerce
. E-
ecommerce is the online trading of physical and digital goods. Some of
the largest companies in the digital economy (e.g., Amazon, Alibaba,
and eBay) are in the business of e-commerce. We will come back to e-
commerce in ► Chap. 17 and explore it in more detail there. One
important aspect of e-commerce is user feedback and
recommendations. Since e-commerce offers consumers the ability to
touch, feel, and test the merchandise to only a limited degree, feedback
and comments from other consumers may add trust to the shopping
experience as explained in ► Box 1.1.

Box 1.1 Recommendations as Market Feedback

Recommendation and consumer feedback are key features in the
digital economy. This often generates strong network effects
resulting in more customers and more sales and may, sometimes,
even lead to de facto monopoly business.
Feedback from consumers can either be gathered directly from
the consumer or indirectly by observing consumer behavior. One
example of direct user feedback is when a consumer is asked to rate
or comment on a product or service after purchase. This information,
together with other information the retailer may gather about the
consumer, can then be used to increase the sales by recommending
other products to the consumer (see ◘ Fig.  1.4). Harvesting
information about the consumers, both directly and indirectly, is a
core activity in the business model of many companies in the digital
economy.

Fig. 1.4 Recommendation system. (Authors’ own figure)

Reviews and feedback also serve another important role: they

build trust in the shopping experience since the customer can read
 reviews and feedback from other customers. Such feedback helps
remove information asymmetries in the digital economy. Encouraged
by feedback and reviews from other customers, potential new
customers will be better informed regarding the product they are
about to buy.
An example of indirect user feedback is when Google is used for
web browsing. Google may then use the search results to build up a
user profile so that more and more accurate advertisements can be
directed toward the user. This increases the value of Google as a
provider of advertisements. The importance of indirect feedback is
even more evident for Facebook where use of the “like” button,
analysis of texts produced by the user, network of friends, and user
activities are run through algorithms to build an accurate picture of
the personality of the user. When you buy books from Amazon, they
also recommend other books that may interest you. This is based on
information stored about books you have previously bought or
shown an interest in and on books bought by other customers with
apparently similar interests.

1.3 Digitization of the Economy

The digital economy is triggered by three technological evolutions:
digitization of data, development of digital ICT infrastructures, and
digital processing and storage. The technological evolutions have
experienced significant breakthroughs and growth in performance and
user adoption in recent decades. ◘ Figure 1.5 shows how these
technological evolutions are related.
Fig. 1.5 Digitization of data, infrastructures, processing, and storage. (Authors’ own figure)

Data has historically been produced in analog formats such as

books, letters, documents, photographs, tape recordings, and videos
recordings. Today an increasing amount of data is produced and stored
digitally. Digitization of data means that the data is encoded as a
sequence of bits (“0” or “1”). Examples of digital data are music stored
as files on a computer, books downloaded on a personal computer or a
tablet, bank account information in an e-bank application, e-mails,
movies and music streamed from the Internet, apps installed on
smartphones, and instant messaging services. Examples of analog and
digital data are shown in ◘ Fig. 1.6. Most telephone services are also
digitally encoded; that is, the voice signal is encoded as a sequence of
digital bits before these bits are transported over a digital ICT
infrastructure to the receiver where the bit sequence is again
transformed into an analog voice signal. Cable television, and, to some
extent, radio (e.g., DAB), is also digitally coded and transported over
digital communication networks.
Fig. 1.6 Analog and digital data. (Authors’ own figure)

Fifty years ago, all data was stored in analog formats, and
digitization of voice and video signals and experiments with
transmission of digital data had just begun. As ◘ Fig. 1.7 shows, the
amount of data stored digitally worldwide has grown from about 1% in
1986 to about 94% in 2007 and that the digitization of the
telecommunication networks has increased from about 20% to 99%
over the same period of time (Hilbert & López, 2011). The evolution
from an analog to a digital society has then taken place over just one
generation.

Fig. 1.7 Evolution of digital data storage and communication. (Authors’ own figure)

Today, an enormous amount of digital data is generated (Data is

giving rise to a new economy, 2017). In 2017, for every minute that
passed, 400 hours of video were uploaded on YouTube, 15 million text
messages were sent between mobile users, three million search queries
were handled by Google, 510,000 comments were posted on Facebook,
and over 45,000 pictures were posted on Instagram. For every second
that passed, more than 10,000 photos were shared on Snapchat
(Schultz, 2019)
. And this is just a small fraction of the total data
generated today. Data is generated directly not only by users but also by
sensor devices connected to the Internet (e.g., weather sensors,
wearable devices, and smart watches). These devices contain
processing and data storage units and are becoming gradually cheaper
and smaller and are rapidly finding new applications producing more
data. In addition, huge amounts of historical analog data—letters,
books, films, television shows, pictures, church registers, and other
documents—are converted into digital data.
This amount of data vastly surpasses what any human mind can
process. Data has become abundant, whereas the human capability to
process this data is a scarce resource. Since all this data is in a digital
format, the data can be transmitted, processed, and stored extremely
efficiently and at low costs, producing a collective knowledge base most
people will have access to. Companies and governments apply big data
techniques and artificial intelligence to analyze these data for various
purposes.
Since the 1980s, there has been an extensive rollout of a worldwide
digital ICT infrastructure. In 2020, the core element of this
infrastructure is the Internet and consists of components such as local
wireless networks, digital subscriber lines, public mobile networks,
satellite networks, optical core networks, submarine optical cables, and
dedicated networks for distress communications, air traffic control, and
broadcast. The digital infrastructure has enabled worldwide
communication of digital data with high capacity, low latency, and high
reliability. The historical evolution of the most important of these
technologies is examined in ► Chap. 2.
It is the combination of digitized data, fast communication
networks, and mass storage that empowers the digitization of the
economy. The business plan of, for example, Facebook requires that all
data (text, images, and video) exist in digital format, that there is a
worldwide digital communication network attaching the users to the
platform, and that there is enough mass storage capacities to store all
user data.
The advances in microprocessors and mass storage of digital data
have resulted in the evolution of digital devices with fast processing
and cheap storage capabilities. Today’s smartphones have the same
processing power as supercomputers had 20 years ago. While 20
megabytes was the standard storage capacity of home computers in
1995, the storage capacity of smartphones 20 years later is more than
100 gigabytes, that is, 5000 times bigger. The cost of storage has shown
similar trends. The cost of one gigabyte of storage in 1995 was about
$1000. Twenty years later, in 2015, the same amount of storage cost
$0.02. This evolution is a result of the packing density of transistors on
microchips which has doubled every 2 years since the early 1970s
(Moore’s law). The quality and inflation-adjusted price of information
technology equipment has decreased on average by 16% per year from
1959 to 2009 (Nambiar & Poess, 2011). Moreover, network capacity
increases and cost of optical fibers decreases following similar laws.
Currently, there is no end to the increase of data produced, stored,
communicated, and processed. The most significant increase is
expected to arise from IoT applications. By 2020, it is estimated that
there will be more than 50 billion IoT devices connected to the Internet
—more than four times as many as in 2017 (Nordrum, 2016).

1.4 Digital Economics

Digital economics is the branch of economics studying digital goods and
services. Hereafter, the term “digital goods” is used when collectively
referring to both “digital goods and services.” This is because there is
seldom a need to make the distinction between “digital goods” and
“digital services.” The full term is used when this distinction is
necessary.
Digital goods comprise everything that is digital: data produced by
users, digital applications and services provided over the Internet, and
the storage and processing of such data. A digital good can be:
– Any kind of software
– Any type of file stored digitally
– A smartphone app and associated services
– Any type of digital information
– Content of a website
– Any communication session
– Any application supported by the Internet
– Trade and bank transactions

Automation
Automation is the use of technology to execute a process without
human intervention. Information and communication technologies
provide significant opportunities for automation, primarily due to the
massive increase in computation, storage capacities, and
communication bandwidth. This has further led to many digital
innovations such as cloud computing, machine learning, big data, and
Internet of Things (IoT). Automation seeks to reduce costs and
increase revenues, competitiveness, and customer satisfaction. It is a
part of the ongoing digitization of the economy and society at large.
ICT-based automation projects can be found in almost every business
sector replacing human workforce by automated digital services,
robots, and algorithms. A big concern for politicians and policy
makers is whether this wave of automation will destroy more jobs
than it creates.
– Financial Technologies (FinTech) is the use of new and emerging
technologies to replace traditional financial services. Examples
include mobile payment services, cryptocurrencies, blockchain
methods, crowdfunding, and smart contracts. A major challenge
associated with FinTech is data security.
– Robotic Process Automation (RPA) is a form of business process
automation based on using software robots as workers instead of
humans. These robots can perform repetitive tasks such as
updating websites, answering questions from customer, and
sending standardized e-mails. More advanced robots, based on
cognitive automation, are being developed.
– In the manufacturing industry, automation is collectively referred
to as Industry 4.0. This implies the use of ICT to create smart
factories based on industrial robots. 3D printing is an example of
an Industry 4.0 technology. Here, industrial products are created on
 demand at customer premises outside the traditional
manufacturing plants.
– Intelligent Transport Systems (ITS) is the use of ICT to automate the
road and transportation sector. Examples of ITS applications
include smart traffic signs and self-driving cars; both contribute to
lower costs of transportation and increased traffic safety.

The complement to digital goods and services is physical goods and

non-digital services, comprising goods that can be touched and felt and
have a physical presence (e.g., cars, books, computers, and furniture)
and services that are not digital in nature (e.g., hairdressing, carpentry,
and teaching).

Definition 1.2
Digital economics is the branch of economics studying digital
services and goods.

The academic field of digital economics overlaps and relates to other

fields of economics. Digital economics is also known under different
designations, each designation having a slightly different focus and
scope. Some of these are:
– Network economy focuses on businesses in which much of the
economic value is generated by network effects (see ► Chap. 9).
Network effects are abundant in the digital economy and explain
how value is generated in several, but not all, digital businesses.
– Platform economy focuses on businesses that act as platforms. The
primary business idea is to connect two or more user groups (two-
sided or multisided markets). Platform economy is closely related to
network economies since network effects are also important drivers
for platform businesses. However, not all network economies are
platform economies and vice versa. Multisided platforms are
described in ► Chap. 10.
– Information economy focuses on information products and how
they are produced and traded. The information economy is part of
the digital economy, but the latter is broader in scope since it also
includes more than pure information goods.
– Data economy focuses on the business of harvesting and analyzing
data. Data is gathered from users or the environment and stored in
large databases. Big Data techniques and artificial intelligence are
applied to analyze the data, where the purpose is to extract
information of value to businesses or governments. Such data may
also be traded on the market, for example, as input to statistics or as
the basis for producing targeted advertisements.
– Virtual economy is the economy of virtual worlds, e.g., World of
Warcraft
and Second Life. To some extent, virtual economies reflect
the real economy regarding the supply and demand of goods, trading,
and network feedback. Virtual economies are mostly disconnected
from the real economy. However, there are examples of virtual
economies that can generate trade in the real economy (e.g., gold
farming in World of Warcraft).
– Internet
economy
comprises the economics of Internet goods and
services. Since most of the economic activity within the context of
digital economics is performed over the Internet, Internet economy
is close in scope to digital economics. One important digital market
that is excluded in the Internet economy is the economics of
telecommunications, that is, the market for broadcast, Internet, and
mobile and fixed network services.
– Attention economy is related to the value created by people’s
attention. User attention is an important element in many digital
business models. The basis for the attention economy is that data has
become abundant, while people’s attention-span remains a scarce
resource. There are business models that exploit people’s attention-
span to generate revenue; the most well-known are those based on
advertisements.
– Sharing economy is the economy in which people or organizations
share goods and services such as Airbnb and Uber. The sharing
economy has also been termed access economy
, peer-economy,
collaborative economy, and crowdsourcing
capitalism.
– Abundance economy is the economy of goods and services that are
abundant; that is, they are close to unlimited in supply. Many digital
services exhibit abundance features, since they can be copied with
zero marginal cost. This challenges one of the most fundamental
 assumptions in neo-classical economics, namely, that resources are
scarce. In several digital economies they are not!
– Digital economics, as defined in this book, encompasses all or parts
of the terms explained above. It is important to point out that digital
economics is a young academic field of study. New terms appears,
and definitions of existing terms are revised as researchers gain
increased understanding of the field and as new technologies expand
the boundaries of the digital economy and enable new business
opportunities.

1.5 Conclusions
In 2020, seven of the ten largest corporations by market capitalization
are in the digital economy business. Moreover, Apple is now about four
times bigger than the corporation that is number eight in the list. Ten
years ago, there were only two companies in the digital economy
among the ten largest corporations (Apple and Microsoft).
The evolution in favor of businesses in the digital economy is a
result of the digitization of society: by 2020, almost all communication
infrastructure worldwide is digital, and almost all data is available in
digital formats.
The companies in the digital economy comprise several industrial
sectors, for example:
– Information and content producers and providers (e.g., newspapers,
television channels, bloggers, YouTube, Netflix)
– Social media providers (e.g., Facebook, LinkedIn, Twitter)
– Manufacturers of devices and software (Microsoft, Apple)
– Operators of fixed and mobile network infrastructure (Internet and
mobile network providers, Starlink)
– Electronic payment services and banking (PayPal, Bitcoin minters)
– E-commerce (eBay, Amazon)
– Sharing service providers (Uber, Airbnb)
– Cloud computing
– Multiplayer interactive game providers
The digital economy has created new business opportunities that
did not exist before as is evident from the list.
Questions
1.
Name the ten largest companies in the world by market  
capitalization. Which of them are in the digital economy?
2.
How large share of the world population has access to the  
Internet?
3.
Does everyone in the world have access to mobile telephony?  
4.
How will digitization of the economy impact the country you live  
in? Search the web for news articles (two or three articles are
sufficient), and discuss how digitization influences the economy,
the public sector, and business domains.

Answers
1.
Apple, Microsoft, Amazon, Alphabet, Facebook, Tencent, Tesla,  
Alibaba Group, TSMC, and Berkshire Hathaway. All of them are
businesses in the digital economy except Tesla, TSMC, and
Berkshire Hathaway.
2.
About 50%.  
3.
No, many people in poor parts of the world do not have access to  
mobile telephony.
4.
The answer to this question depends on where you live.  
However, there are a few general observations such as access to
the internet and mobile services, access to vast amounts of
information, automation of industrial production, automated
access to societal functions, and so on.

References
Christensen, C. (1997). The innovators dilemma: When new technologies cause great firms to fail.
Harvard Business School Press.

Data is giving rise to a new economy. (2017, May 6). The Economist.

Grossman, R. (2016, March 21). The industries that are being disrupted the most by digital.
Harvard Business Review.

Hilbert, M., & López, P. (2011, April). The world’s technological capacity to store, communicate,
and compute information. Science, 332.

ITU statistics, 2018.

Nambiar, M., & Poess, M. (2011). Transaction performance vs. Moore’s Law: A trend. In
Performance evaluation, measurements and characterization of complex systems (TPCTC 2010)
(Lecture Notes in Computer Science) (Vol. 6417). Springer.
[Crossref]

Nordrum, A. (2016, August 18). Popular internet of things forecast of 50 billion devices by 2020
is outdated. IEEE Spectrum.

Reid, P. (2015, October 27). IFPI Digital Music Report 2015: Global digital revenues match
physical for first time. The Music Network.

Schultz, J. (2019, August 6). How much data is created on the internet every day. Micro Focus
Blog.

Tapscot, D., Ticoll, D., & Herman, D. (2006). Digital conglomerates: Setting the agenda for
enterprise 2.0. New Paradigm Learning Corporation.

Vincent, J. (2016, July 4). UN condemns Internet access disruption as human rights violation.
The Verge.

Wikipedia. (2021) List of public corporations by market capitalization. (Wikipedia website

visited April 12, 2021)

Further Reading
Brian Arthur, W. (2015). Complexity and the economy. Oxford University Press.

McAfee, A., & Brynjolfsson, E. (2017). Machine, platform, crowd. W. W. Norton & Company.

Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to network economy.
Harvard Business School Press.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_2

2. Information and Communication

Technologies
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Transistor – Internet – Cellular mobile networks – World

Wide Web – Internet of Things

Learning Objectives
After completing this chapter, you should be able to:
– Explain how information and communication technology has
evolved toward increasingly complex systems and
applications.
– Understand that the ICT evolution has followed three
intimately coupled lines of development: evolution of basic
hardware technologies, evolution of fixed and mobile data
networks, and evolution of application protocols, software,
and services.
– Understand the impacts and potentials of the Internet of
Things evolution.

2.1 Introduction
The evolution of information and communication technologies has
followed three parallel timelines:
– The innovation
of technologies from simple telephone and telegraph
systems to the Internet supporting social media, sensor networks,
apps, and many other digital services.
– The convergence
of services in which the telephone and telegraph
networks are replaced by the Internet (► Chap. 3)
.
– The evolution of the telecommunications business itself from
monopoly to competitive markets (► Chap. 5).
In this chapter, we will look more closely at the technological
evolution and postpone the other items to later chapters as indicated
above.
The evolution that eventually led to today’s digital economy started
with the invention of the transistor in 1947 (see ► Box 2.1). This
device could be more densely packed, was much cheaper, used less
energy, was easier to handle, and was much more reliable than vacuum
tubes. The transistor, and the miniaturization capabilities it eventually
offered, led to a technical evolution at a speed the world never had seen
before.
In 2019, approximately 23 billion devices containing microchip
CPUs (central processing units) were connected to the Internet, making
the global information and communications technology (ICT)
infrastructure the largest machine ever built. Moreover, the number of
connections increases by 10% per year, corresponding to a doubling
time of 7 years. The forecast for 2025 is that more than 40 billion
devices will be connected to the Internet (Cisco Visual Networking
Index, 2019).

Box 2.1 The Transistor

The transistor is a semiconductor device used to switch and amplify
electronic signals. It effectively replaced vacuum tube technology and
enabled the production of cheap, low-power, and small electronic
devices. It is the basis for almost all ICT devices such as
microprocessors, personal computers, smartphones, and other
electronic tools. It is the most important invention of the twentieth
century, perhaps the most important invention of all times.
 Julius Edgar Lilienfeld had already filed a patent for the field
effect transistor in 1925. However, because of the lack of high-quality
semiconductor materials, it was impossible to build a working
transistor at that time. The first practical implementation of the
transistor was done at Bell Labs, USA, by John Bardeen, Walter
Brattain, and William Shockley in 1947. What they invented was the
first point-contact transistor which they patented the year after. They
received the Nobel Prize in Physics in 1956 for their “research on
semiconductors and their discovery of the transistor effect” (See ►
https://​www.​nobelprize.​org/​prizes/​physics/​1956/​summary/​).
The transistor radio was the first commercial device designed
using transistors (1954). In its early days, the transistor also found
its use in pocket calculators, hearing aids, telecommunication
switching equipment, and then, finally, the personal computer and
the mobile phone. Today, transistors are mostly used as building
blocks for integrated circuits which, in turn, are used to produce
personal computers, smartphones, and other electronic devices. In
2014, more than 1018 transistors were produced. This is more than
100 million transistors for each human being on Earth. From 1960 to
2018, altogether 1.3 × 1022 transistors have been manufactured
(Laws, 2018).
The size of a single transistor has continually gotten smaller since
its inception in the 1950s, quite accurately following Moore’s law.
While the Intel 4004 microprocessor released in 1971 had 2300
transistors, each with a size of 10,000 nanometers, the 22-core Xeon
Broadwell-E5 microprocessor released by Intel in 2016 has
7,200,000,000 transistors each with a size of 14 nanometers
(Wikipedia, 2020). More transistors mean in general more
computing power. Whether or not the size of transistors can be
further reduced in the future according to Moore’s law is an open
issue. In the end, quantum effects, heating, and thermal noise may
limit the minimum size of a transistor.

2.2 Timeline for the ICT Evolution

◘ Figure 2.1 shows the timeline for selected innovations that were
essential for the development of the digital economy and the year they
became commercially available or reached the mass market. The
technologies listed in ◘ Fig. 2.1 are categorized as hardware,
mobile/wireless, or software/services.

Fig. 2.1 Timeline of ICT innovations. (Authors’ own figure)

Prior to the commercialization of the World Wide Web in 1993, the

telecommunications technology and the information technology were
developed along two different paths.
1.
Telecommunications is an old technology branch going back to the  
first electronic telegraph systems developed in the early nineteenth
century. The predecessor of the International Telecommunications
Union (ITU) was established in 1865 with the task to standardize
the telegraph technology and encourage the establishment of a
global telegraph network. Since then, specification and design of
telecommunications networks and services have been done by the
ITU and other standardization organization, the
telecommunications industry, and international research programs
in close cooperation. Until 1993, the development was primarily
associated with digital transmission and switching of telephone
services and simple data services.
Information technology is a much younger science with roots back
 Information technology is a much younger science with roots back
2.
to the 1950s when the first transistorized computers were  
designed. It has now become one of the world’s largest and most
influential industries. Basic information technologies such as
computer architecture, processing platforms, data storage,
algorithms, and programming were studied and developed at
universities, by research organizations, and in the computer
industry. In 1993, there were still only few applications of the
information technology that required extensive support of
communication technologies, the most important of which was e-
mail on the Internet.
All this changed in 1993 when the two fields of information
technology and telecommunications merged as a single technology now
known as information and communication technology (ICT). The event
that led to this transformation was commercialization of the World
Wide Web in 1993, soon leading to a massive requirement for remote
interactions between computers and for distributed processing of
information. Since 1993, information technology cannot exist without
telecommunications and vice versa.
The World Wide Web is enabled by the following technologies:
Uniform Resource Locator (URL
) for accurate addressing of
information packages on the web, Hypertext Transfer Protocol (HTTP
)
for communication with web pages, and Hypertext Markup Language
(HTML
) for formatting and writing web pages. Together, these
technologies allow users to post and access documents, images, videos,
and other information across the Internet. Mosaic (1993) was the first
graphical web browser which contributed to popularizing the WWW
and, consequently, the Internet itself. Mosaic was later followed by
Netscape, Internet Explorer, and Google Chrome.
The Internet is a global system for interconnected computer
networks based on technologies and protocols such as Ethernet (1974)
and TCP/IP
(1974), allowing data to be transferred between two or
more computers. The TCP/IP protocol suite was developed and tested
as a part of the ARPANET project financed by the US Department of
Defense but is now the worldwide open standard for data transmission
on the Internet (see ► Box 2.2).
 Optical fibers were invented in 1965 and, as the technology
matured during the 1980s, eventually provided a high-speed global ICT
infrastructure for the Internet. Most of the Internet backbone network
is built using optical fibers. A single optical fiber, which is thinner than a
human hair, can carry several hundred terabytes of data per second. An
optical cable, consisting of several (sometimes hundreds of) optical
fibers, can accommodate all traffic generated on the Internet today.
With optical fiber technology, the Internet can be built with abundant
capacity for decades to come.
One of the first “killer applications” of the Internet was e-mail which
was standardized in 1982 with the Simple Mail Transfer Protocol
(SMTP
). E-mail soon became the key technology for exchange of
messages.
The first commercially available microprocessor was the Intel 4004,
released in 1971. It was based on the transistor technology that had
been commercialized two decades earlier, enabling reliable and low-
cost digital computing. Today, microprocessors are found in everything
from computers to smartphones, refrigerators, cars, and toys. The
personal computer (PC
) was developed in the early 1970s, but it did
not reach the mass market until 1977 with the release of Apple II and
Commodore PET.
The PC disrupted the existing time-sharing mainframe and
minicomputer systems by offering a dedicated low-cost multipurpose
computing device for end users. The Microsoft Disk Operating System
(MS-DOS) was released in 1981. It provided the technological basis for
Microsoft’s later products and dominance in the digital economy. The
first laptop available for the mass market was the Toshiba T1100,
released in 1985. The laptop is a PC combining display, keyboard, input-
output devices, and storage in a miniaturized package. In 2018, more
than 160 million laptops were sold (Shipment forecast of tablets,
laptops and desktop PCs worldwide from 2010 to 2024. Statista, 2020).
However, this number is small compared to the over 1.5 billion
smartphones sold the same year (Number of smartphones sold
worldwide from 2007 to 2021. Statista, 2020).
The Nordic Mobile Telephone (NMT), released in 1981, was the
world’s first automatic cellular mobile telephony system. It introduced
automatic roaming and handover but supported only voice
communication. The Global System for Mobile Communications (GSM)
—supporting voice, messaging, and data services—pioneered digital
radio communication (1991). Initially, data communication was slow
(less than 10 kilobits per second) and ineffective, but the data rate was
later enhanced by technologies such as GPRS and EDGE. High-speed
data, particularly on the link from base station to mobile terminal
(usually referred to as the downlink), was introduced in 3G mobile
systems (2003). The broadband technology was developed further in
the 4G mobile systems launched in 2009. The latest mobile technology,
5G, was used at the 2018 Winter Olympics in South Korea and is now
(2020) being rolled out on a large scale in Europe, Asia, and North
America. 5G will support the Internet of Things (IoT). The family tree
for mobile networks is contained in ► Box 2.3.
Bluetooth (1997) interconnects devices over very short distances,
while Wi-Fi (1997) offers local network access using packet radio. Both
technologies are widely employed today. Android (2007) and iPhone
(2007) transformed smartphones into advanced computers also
supporting the app ecosystems (Apple’s App Store and Google Play).
The smartphone apps are the main building blocks of the emerging
location-based services such as Airbnb (2008) and Uber (2009), both
being sharing services. Google (1998) enables users to search the
World Wide Web for information. When Google was launched, there
were several competing search engines; however, Google turned out to
be technologically superior to its competitors and captured most of the
market. The initial business of Netflix (1998) was DVD rental and sales
by mail; now, Netflix provides online media streaming services to
subscribers in over 190 countries. ► Amazon.​com (1997) launched a
successful e-commerce website.
When home computers and laptops became ubiquitous in the
developed world, Internet services such as Wikipedia (2001), Skype
(2003), Facebook (2004), YouTube (2005), Twitter (2006), Dropbox
(2007), and Spotify (2008) emerged. These services, satisfying different
user needs, were not the first in their respective service category.
However, with a combination of technological superiority, smart design,
and luck, they managed to become the dominating services. Every year
new innovations emerge in different markets and fields. Bitcoin
(2009)
emerged with the idea of disrupting the banking system by offering a
cryptocurrency that enabled trade and money transfers without third-
party involvement. Likewise, Snapchat (2011) sought to fill gaps in the
social media market by enabling private sharing of pictures and videos
on mobile devices.

Box 2.2 The ARPANET

The ARPANET
was a project with the major goal of building and
demonstrating a data communication network based on packet
switching. It was also the first communication network to implement
the TCP/IP
protocols. The ARPANET was funded by the US
Department of Defense and launched in 1966. Packet switching was a
novel technology at that time, challenging the established circuit
switching technique used in telephone networks. The two key
advantages of packet switching over circuit switching were efficient
resource sharing and resilience against node and link failures. Some
scientists and engineers doubted packet switching could be
implemented due to its complexity.
In 1969, the ARPANET project built an experimental packet
switched network connecting a few computer sites. In subsequent
years, the ARPANET was refined and expanded to the network shown
in ◘ Fig.  2.2. The first international connection in the ARPANET was
to Norway via a satellite link in 1973.
Fig. 2.2 The ARPANET in 1974. Note that the first node outside the US was at Kjeller in
Norway. (Public domain, source: ► https://​commons.​wikimedia.​org/​wiki/​File:​Arpanet_​
1974.​svg)

ARPANET was the predecessor of the Internet where the key

technologies in the current Internet were developed and tested. This
includes packet switching, protocol layering, and the TCP/IP
protocol
suite. Many of the early services of the Internet, such as e-mail and
file transfers, were also first developed and tested on the ARPANET.
The ARPANET was decommissioned and replaced by NSFNET in
1990 and became the first part of the current Internet.
The early Internet was mainly used at universities and research
establishments. The network was hardly known outside these circles
until the World Wide Web was commercialized and taken into use by
several telecommunications carriers in 1993 and, thereby, becoming
available to the public. The WWW technology had been invented by
Tim Berners-Lee at CERN already in 1989, but since it was not
invented by the telecommunications industry, it took a long time
until they discovered the potential the new technology would have
for the data communication market, a market the carriers had strived
to build up for more than a decade without succeeding.
Box 2.3 Family Tree of Mobile Networks
Altogether five generations of mobile systems have been developed:
First generation (from 1981): NMT (Nordic countries), TACS
(UK), Radiocom2000 (France), and C-Netz (Germany) offering only
analogue telephony. These systems supported primitive roaming
capabilities, though the method used in NMT became the basis for
the more sophisticated roaming capabilities of GSM.
Second generation (1991): GSM offering digital telephony, data
communication at speeds up to 10 kilobits per second (kbps), and
short message service (SMS) over signaling channels. GSM was
designed for automatic international roaming and non-disruptive
handover.
Third generation (2001): 3G (or UMTS) is a dual system
offering packet radio services at a 128 kbps (initially) for Interne
services and GSM services for telephony and SMS. The architecture
consists of two separate network architectures for data and
telephony but using the same radio interface based on spread
spectrum technologies. 3G is an extension of both the Internet and
the telephone network.
Fourth generation (2009): 4G is an extension of the Internet
offering only packet radio services including voice over IP (VoIP),
narrowband data, broadband data, and streaming services over a
dynamic mix of narrowband and wideband data channels.
Interconnection with the fixed telephone network is via conversion
units at the interface between the telephone network and the 4G
network.
Fifth generation (2018): 5G is based on 4G but offers new
features such as very high data rates, edge computing (cloud
computing close to the mobile user, e.g., in the base station, to reduce
latency), network slicing (allowing independent providers to operate
simultaneously over the same infrastructure offering complex
services to the same user), and connection of millions of remote
sensors and other devices. 5G will be one of the basic technologies of
the Internet of Things.
Intermediate technologies exist between the generations such as
GPRS (General Packet Radio Service) offering packet radio with
 increased data rates over GSM and HSPA (high-speed data access) for
increased data rates over 3G.

2.3 Factors Constraining Evolution

Digital services cannot exist without ICT. Innovations in computer
networking and wireless technologies give rise to new services which,
in turn, have major impacts on the digital economy at large. Note that
some of the basic information and communication technologies still in
use, including TCP/IP
, HTTP (World Wide Web), Ethernet, and GSM,
are now more than 25 years old. Several of these technologies have
been expanded and improved several times during their lifetime. For
example, IP exists in two versions (IPv4 and IPv6). GSM has generated a
whole new family of mobile communication systems—3G, 4G, and 5G—
all of them built on the basic principles first outlined in the GSM
project. On the other hand, the basic connection-oriented protocol on
the Internet—TCP—has been unchanged since 1974.
Despite all these improvements, the original technologies are still
widely used. For example, IPv4 (42 years old) and GSM (30 years old)
are still important Internet and mobile network technologies,
respectively. Now, 4G and 5G technologies are rapidly replacing GSM
but, still, mobile networks support GSM to back global roaming
services. Any efforts to shut down GSM have, so far, failed except in a
few countries (e.g., Australia, USA, and Singapore).
Communication technologies are evolving slowly. The most
important reason for the slow adaptation of some of the new
technologies is the huge investments required for implementing them.
Even a small improvement of a technology is expensive to install,
simply because of the vast volume of existing equipment designed to
the old standard. Therefore, it may sometimes take more than 10 years
before the technology is taken into use after it is specified and ready for
the market. For example, it took more than 10 years from the HTTPS
specification (encrypted web access) was finalized until it was taken
into use. IPv6 was ready for implementation in 1996. Nevertheless, in
2016, more than 95% of the Internet traffic was still carried on IPv4
networks. This is because Network Address Translation (NAT) has
increased the available address space for IPv4 and, hence, postponed
the introduction of IPv6. Because 4G and 5G mobile networks only
support IPv6 and IPv6 is implemented on almost all input/output
devices of computers, this is now (2020) about to change. It is assumed
that IPv6 will soon replace most of IPv4 worldwide.
On the other hand, both the development time and adaptation time
for many app-based digital services (such as Airbnb and Uber) are very
short. The reason is that many of them are simple software packages—
easy to develop, install, and use. The rapid evolution of apps took place
after iPhones and Android phones were marketed in 2008.
One important requirement for introducing a new technology is
backward compatibility; that is, the new technology should support
equipment or software designed to the old standard. One compatibility
requirement is that new equipment should be capable of operating in
the old environment. This objective is fulfilled for public mobile
communication; a smartphone designed for 5G must also support 4G,
3G, and GSM so that it can be used everywhere. This implies that the
smartphone must support the radio interface for all mobile standards
so long as these standards are in use. In addition, it must support Wi-Fi
and Bluetooth. This backward compatibility ensures that new families
of mobile systems can be introduced smoothly without rebuilding the
network completely.
The reliability of the infrastructure is also a concern that may
influence the rate by which new technologies are implemented. ICT and
electric power production are the two most critical infrastructures of
society. If any of these infrastructures stops, society will soon grind to a
halt since all other infrastructures depend critically on them. All
activities of modern society involve computation, sharing, and storage
of data. Therefore, these activities are vulnerable to cyberattacks where
the purpose is to destroy, disable, or gain illegal access to computer
resources and infrastructures. Events have also shown that most ICT
systems are taken out of service either at the same time a power outage
occurs or shortly afterward if devices or networks are equipped with
standby power such as batteries or diesel aggregates (which most of
them are not) (Northeast blackout of 2003, 2020).
Innovations in ICT will continue to have impact on the digital
economy in the future. Technologies such as machine learning, robotics,
smart factories, smart cities, and 3D printing all show great potential
for disrupting existing business sectors and providing the foundations
for upcoming digital services. Machine learning techniques are already
utilized in several digital services. One example is algorithms used to
recommend products to consumers based on previous customer habits.
Another example is voice recognition systems such as Apple’s Siri. This
evolution is considered further in ► Chap. 20.

Box 2.4 Cryptocurrencies

Cryptocurrency is a type of digital currency that uses cryptography to
secure transactions and generate new units. Transactions in
cryptocurrencies are performed without any centralized authority.
Cryptocurrency uses the blockchain technology employing a public
and decentralized database of encrypted records. Blockchain is a
versatile technology that can be used to design smart contracts as
well as cryptocurrencies. A general advantage of cryptocurrencies is
their ability to support secure transactions of money without a third-
party stakeholder involved, potentially with reduced cost and lower
transaction delay.
Bitcoin
(2009) was the first decentralized cryptocurrency and the
first practical implementation of the blockchain. Bitcoin involves
many distributed miners that confirm transactions between two
parties. Transactions are recorded as a chain of blocks (thereby the
name “blockchain”) on a decentralized and distributed public ledger.
Bitcoin has the potential to become a true global digital currency.
It can be traded on international digital currency exchanges.
However, the price of Bitcoin has been extremely volatile compared
to other currencies such as US dollar and euro (see ◘ Fig.  2.3). The
current value of 1 Bitcoin is about 60,000 USD (April 2021), rising
from about 7000 USD in April 2020. Because of this, the use of
Bitcoin in the trading of goods and services has been limited. The
popularity of Bitcoin has also spurred the creation of many other
cryptocurrencies such as Ethereum
, Litecoin
, Ripple
, and IOTA
.
There are currently more than 3400 different cryptocurrencies in
various stages of development.
 Fig. 2.3 The price of one BTC in US dollars from January 2014 to April 2021. (Authors’ own
figure)

Cryptocurrencies, in particular Bitcoin, have met much criticism

and have been compared to economic bubbles such as the Tulip
Mania in 1637 and the South Sea Bubble in 1720. Cryptocurrencies
have also been used in money laundering and in funding of criminal
activities.

2.4 Internet of Things

Interactions between humans (e.g., speech and e-mail) and between
humans and machines (e.g., downloading of films and payment
services) are traditionally the core elements in the evolution of
telecommunications technology comprising network architecture,
network management, communication protocols, and services. Most
telecommunications statistics is still concerned with the number of
individuals connected to the Internet and mobile networks, the amount
of traffic they generate, and the revenues this traffic creates for the
telecommunication operators. This simple picture of
telecommunications is now changing. During the last decennium,
autonomous communication between machines not involving humans
has become increasingly important. This includes remote sensing and
control, grid computing, driverless cars, smart homes, smart cities,
smart grids, infrastructure management, and several other
applications. IoT devices may report to and receive commands from
centralized platforms, for example, automatic reading of power and
freshwater consumption, or interchange message with one another
without involving a central processing platform, for example, vehicle-
to-vehicle communication and vehicle-to-road infrastructure
communication for autonomic driving.
Cisco has estimated that there will be almost 30 billion devices
connected to the Internet by 2022. About 50% of these devices will be
IoT devices and 25% will be smartphones (Cisco Annual Internet
Report, 2020). The remainder will be TVs, PCs, and tablets. On the other
hand, IoT devices will only generate 6% of the global Internet traffic.
This is so because most IoT devices transmit few and short messages as
compared to other users of the Internet. According to Business Insider,
there are almost three times as many smart devices in people’s homes
as shown in Cisco’s statistics; however, only a few of them are
connected to the Internet. If these devices are also connected, then the
increase in IoT devices will be dramatic, resulting in a tripling of the
estimates of Cisco (Patel, 2018).
The applications of IoT devices are commonly divided into five
categories: consumer, commercial, industrial, infrastructure, and
military. Several applications exist already, but the number of
applications is expected to escalate dramatically during the next few
years. One consequence is that IP version 6 must be built out on a large
scale to provide enough addressing space for the new applications. The
5G mobile network is already designed for this evolution, and it is
assumed that 5G will play an important role in the evolution of IoT.
Consumer applications include smart home applications, elder care,
and medical supervision managed by the users. Smart home
applications may include wireless communication with light switches
and appliances and remote control of air conditioning and security
systems using smartphones. Assistance services includes voice-
controlled devices (e.g., light switches and door openers), medical
monitoring, activity monitoring, and other services making life easier
for elderly and disabled people.
Commercial applications include smart health care, building and
home automation, and smart transport systems. These applications are
based on a mix of device-to-platform communication and autonomous
communication between devices. Smart health care comprises
technologies for monitoring of patients both at home and in hospitals.
Examples are remote reading of measuring devices (e.g., blood pressure
and heart rate monitors), monitoring and actuating implants such as
pacemakers and insulin pumps, and supervising patients with chronic
diseases and elderly people. Aruba estimates that IoT technology in
health care will increase the workforce productivity and reduce the cost
of health care by more than 50% (State of IoT Healthcare, 2019).
Building and home automation include safety alarms, energy saving,
comfort, and occupancy monitoring (e.g., reducing heating if no one is
at home).
Examples of smart transport systems are communication between
vehicles, between vehicles and road infrastructure, smart parking, toll
collection systems, fleet management, and road assistance (e.g.,
automatic accident alarm generated automatically by the vehicle). One
application that has taken off at an unexpected rate is renting of electric
kick scooters. The autonomous system combines services such as
location management, payment, and status reporting.
Industrial applications facilitate the evolution of existing industrial
automation using big data, AI, robotics, and autonomous interactions
between devices, machines, control systems, and management systems
at a much larger scale than today. Networked sensors and complex
algorithms analyzing and acting upon real-time measurements will
improve safety management, increase process efficiency, reduce waste
of time and resources, and admit real-time plant optimization. This is
often referred to as Industry 4.0 or the Fourth Industrial Revolution.
Infrastructure applications are management and control of
infrastructures such as freshwater and sewage systems, railway tracks,
roads, tunnels, bridges, energy production and delivery networks, and
so on. The category also includes environmental monitoring of, for
example, air pollution, freshwater quality, wildlife habitats, and soil
conditions. IoT may also support smooth coordination between
authorities (political, technical, and managerial) since the management
of a particular infrastructures often involves several authorities.
Military applications include technologies for surveillance,
reconnaissance, and battlefield operations.
Several technologies support IoT networking. These are short-range
applications such as Bluetooth, Wi-Fi, Zigbee, and Z-Wave; medium
range systems using 5G mobile networks; long range radio
communication using ultralow bit rates (e.g., 300 bits per second);
satellite networks, in particular, satellite terminals with small antennas
(VSAT); and wired systems such as Ethernet, cable networks, and
powerline communication. Several technologies may be connected in
tandem, for example, a Wi-Fi network connecting several local sensors
to a local processing platform connected to a 5G network.
There are several concerns regarding IoT, the most important of
which are related to privacy, security, and safety. Efficient health-care
applications generate enormous amounts of personal data. These data
may be abused by the authorities for political manipulation and social
control. Other abuses may be surveillance of people and recording of
consumer behavior. In many IoT applications, sensitive information
about people may be sent over multiple hops and stored and processed
several times in different computers, severely increasing the probability
that confidential information is compromised.
Security is the biggest concern related to IoT. The IoT will consist of
billions of devices designed to different security standards (if any at
all). Security concerns include weak authentication, weak or no access
control protection, unencrypted messages, no firewall protection, and
improper security updating. The result is a network with very many
access points vulnerable to cyberattacks. In some applications, for
example, autonomous driving and remote monitoring of medical
devices, cyberattacks may result in accidents and loss of life. Poorly
protected IoT devices may be captured and assimilated in large botnets
to become formidable tools in distributed denial of service attacks.
Internet of Things Security Foundation is a nonprofit international
organization promoting security best practices and management.
Safety is an important issue in IoT systems, particularly, those used
in complex systems where errors may lead to disasters. IoT systems
may, as all software systems, contain bugs, flaws, and unintended
interactions. One characteristic of complex systems is that they may
contain pathways into hidden states causing deadlock or other errors.
These states may be legitimate operational states in some applications
but may cause problems in other cases. The existence of such
conditions is difficult to identify using standard software production
and checking tools. Systems controlling hardware are particularly
vulnerable to errors leading to unsafe physical states because of the
damage they may instigate, for example, power grids, sluices regulating
river flows, pacemakers, or traffic safety systems for autonomous
driving.
The term Internet of Things was coined by Kevin Ashton at Procter
& Gamble in 1999 (Ashton, 2009). At that time, a few simple logistics
applications existed such as routing of luggage in airports, tagging of
containers and goods for simple identification and sorting, and locating
objects and animals (e.g., tagging of cattle and pets). Between 2008 and
2009, the number of objects connected to the Internet surpassed the
number of people connected to the network. According to Cisco, the
ratio of connected objects to connected people was only 0.08 in 2003,
while in 2010 this number had increased to 1.84 (Evans, 2011).
There are several obstacles slowing the evolution of IoT: lack of
killer applications; unclear business propositions and usability; lack of
standards and interoperability; security, safety, and privacy concerns;
and, perhaps the most important, IoT not fitting easily into traditional
public or industrial governance structures.

2.5 Conclusions
The evolution of the digital economy is intimately connected to several
technological innovations.
– The evolution started by the invention of the transistor in 1947. The
evolution ever since is intimately linked to how densely transistors
can be packed on microchips and how fast microelectronic circuits
can operate.
– The Internet offered a simple, cheap, and effective communications
platform for data communication (especially since the early 1980s).
The Internet allowed computers to be interconnected in a dynamic,
flexible, and effective way.
– The World Wide Web created the real killer applications
(commercialized in 1993) enabling social media, high-speed
streaming services, and sharing services. It is the World Wide Web
that created the digital economy as we know it today.
– The evolution of data communication over digital mobile cellular
networks started with GSM in 1991 and gained speed when the 3G
technology was introduced 10 years later. Cellular mobile systems
make communications ubiquitous in a new way by making it
independent of place and time.
These events have led to the convergence of networks and services
as explained in ► Chap. 3. With the Internet of Things, ICT has entered
new and enormous fields of applications.

Questions
1.
Give examples of both new and old information and  
communication technologies that are still in use today.
2.
What are the major challenges that may hamper the evolution of  
IoT?
3.
Why are mobile systems (public or local) so important in IoT?  
Answers
1.
Old technologies include TCP, IP, and SMTP. New technologies  
include smartphones, Bitcoin, and 5G.
2.
Privacy concerns, safety and reliability of operation, prone to  
cyberattacks, lack of standards, does not fit well with traditional
governance standards.
3.
Because very many applications require wireless access (health  
care, autonomous driving, smart transport, environmental
surveillance, etc.); easier and faster to install.
References
Ashton, K. (2009, June 22). That “Internet of Things” Thing. RFID Journal.

Cisco Annual Internet Report (2018–2023). (2020, March 9). White Paper. Cisco.

Cisco Visual Networking Index: Forecasts and Trends 2017–2022. (2019). White paper. Cisco
public.

Evans, D. (2011). The Internet of Things. How the Next Evolution of the Internet is Changing
Everything. CISCO White Paper.

Laws, D. (2018, April 02). 13 sexillions & counting: The long & winding road to the most
frequently manufactured artefact in history. Computer History Museum.

Northeast blackout of 2003. (2020, December 17). Wikipedia.

Patel, H. (2018). How IoT & smart home automation will change the way we live. Tristate
Technology.

State of IoT Healthcare. (2019). Aruba (Hewlett Packard Enterprise).

Wikipedia. (2020, December 17). Transistor count.

Further Reading
Abbate, J. (1999). Inventing the internet. The MIT Press.

Brynjolfsson, E., & Saunders, A. (2013). Wired for innovation. The MIT Press.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_3

3. Convergence of Technologies and

Services
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Dedicated network – Integrated network – Next-generation

network – Mobile Internet – Fixed Internet

Learning Objectives
After completing this chapter, you should be able to:
– Explain how the evolution from dedicated networks for
telephony, data communications, and broadcast to an
integrated network supporting all types of technologies has
taken place.
– Appreciate the roles of the Internet and mobile networks as
generators of this process.
– Understand how the convergence has changed the
technological landscape.

3.1 Dedicated Networks

Historically, separate and dedicated communication networks were
designed for carrying different telecommunication services. This means
that a separate and dedicated physical infrastructure was constructed
for each digital service offered. Hence, the provision of a new service
(e.g., TV or radio) demanded the building of a completely new physical
infrastructure, in which most cases had huge up-front costs. Examples
of such dedicated network include:
– Public switched telephone network (PSTN) for telephone services.
This network also supports the facsimile service and several value
added services, for example, premium rate calls, free-phone,
televoting, and particular call handling services such as time-
dependent call routing.
– Public mobile networks for voice, data, and messaging services (SMS
and MMS).
– Wireless one-way networks for audio and television broadcasting.
– Cable and satellite networks for audio and television broadcasting.
– Telex network for low-speed text transmission.
– Satellite access networks for communication with ships, aircraft, land
mobile terminals, and remote areas (e.g., Inmarsat, Skyphone, and
Iridium).
– Dedicated data networks for various digital data services, for
example, the X.25 network and the Internet.
– Dedicated networks for search, rescue, and distress communications.
– Fixed and wireless telegraphy networks using the International
Morse Code, for example, used for distress communications with
ships and for railway signaling.
All these networks, except the telegraphy networks, the telex
network, and some early data networks (X.25), are still in operation
today. Telegraphy based on the International Morse Code was
standardized for continental Europe in 1851. Other countries, e.g., the
UK and USA, used other incompatible methods. In 1865, the
International Telegraph Union—now the International
Telecommunications Union (ITU)—was established for the purpose of
technical, operational, and commercial standardization of a single
international telegraphy service to be used in all parts of the world. The
service was mandatory for commercial ships for distress
communications until 2000 when the International Maritime
Organization (a specialized UN organization) decided that telegraphy
was no longer needed since satellites offered a more reliable service.
The world’s oldest telecommunications service was terminated after
about 150 years of operation. However, the technology is still used by
radio amateurs. The services offered by the telex network has been
replaced by email, and the public X.25 packet switched data network
has been replaced by the Internet.
Audio broadcasts have been sent as analog signals over wireless
networks since the 1920s (also denoted “over the air audio
broadcasting”). From 1993, digitized audio broadcasts have also been
offered over the Internet (Internet radio, web radio, or webcasting).
Over the air radio broadcasting is now also shifting from analog to
digital technology (e.g., digital audio broadcast (DAB
)) and are also
offered as real-time services over the Internet (Internet radio) as well
as over wireless networks. Television programs have, in addition to
over the air transmission, been broadcasted over dedicated cable
networks since the 1950s and over satellites since 1976. From the early
2000s, television programs have also been available over the Internet.
Some television broadcasters, such as the BBC, offer their television
programs on web pages in addition to delivering them over cable or
satellite networks. The Internet allows listening to or viewing
broadcast services at any time independently of when the program was
transmitted over wireless or other media. There are also several pay-
per-view Internet services such as Netflix and HBO offering movies,
series, and other video content. Some television companies only offer
their programs over the Internet.

3.2 Early Attempts of Service Integration

Attempts to design one integrated network supporting services carried
by the PSTN and dedicated data networks started in the late 1970s
resulting in the Integrated Services Digital Network (ISDN) standard in
1988. The failure of ISDN was that it did not amalgamate the telephone
network and the packet data network into a common network except at
the subscriber line. The transport network for telephony and data still
existed as two separate networks. The next attempt was to merge the
telephone network and packet data network into a single transport
network using the Asynchronous Transfer Mode (ATM) technology.
This attempt failed because the technology was ready for
implementation at the same time as the Internet expanded into a
worldwide platform open for access for everyone. At the same time,
cellular mobile communications grew rapidly offering location-
independent services on a global scale.
To appreciate the challenges associated with service integration,
some basic technological concepts must be understood: circuit
switching, packet switching, connection-oriented transfer of data, and
connectionless transfer of data.

Definition 3.1
Circuit switching implies that a physical two-way (duplex)
connection between the users exists so long as the communication
session lasts.

The telephone network is a circuit switched network where a two-way

physical connection is established over several telephone exchanges for
the duration of the call. A circuit switched connection can be used for
only one communication session at a time.

Definition 3.2
Packet switching implies that chunks of data (e.g., speech or video
samples, emails, web pages, and so on) are sent as independent
packets of data.

At the switching devices in the network (the routers), the packets are
queued before they are forwarded to the next router or to the recipient.
Moreover, packets belonging to different communication sessions are
arbitrarily mixed when they are sent over the communication links
between routers and between routers and terminals. For this reason,
packet switched networks offer better utilization of the communication
infrastructure than circuit switched networks: for voice
communication, each direction of a telephone circuit is on average busy
only 40% of the time so that 60% of the transmission capacity is
wasted. For the transfer of pictures and documents over the telephone
network (telephoto or facsimile), both directions of communication are
used, but only one is occupied by transfer of information.

Definition 3.3
Connection-oriented transfer means that an association is established
between the users before information is transferred between them.
The association is retained until all information has been
transferred.

Definition 3.4
Connectionless transfer means that no such association is
established, and data is transferred in independent packages.

A circuit switched connection is always connection oriented, while a

data connection may either be connection oriented or connectionless.
Connection-oriented packet switched connections are also referred to
as virtual circuits. Note that the Internet protocol (IP) itself is
connectionless. Connection-oriented data transfer is achieved by
imbedding a connection-oriented protocol in the IP packets. The most
used connection-oriented protocol on the Internet is the Transmission
Control Protocol (TCP) (the TCP/IP protocol suit).
The idea behind the Integrated Services Digital Network (ISDN) was
to provide both telephone services and packet switched data
transmission over the infrastructure of the telephone network. It turns
out that it is impossible to incorporate packet switching in a circuit
switched environment. The ISDN, therefore, consists of two parallel
transport networks: the telephone network and the—now obsolete—
X.25 packet switched data network. Therefore, the ISDN never became
a common network for circuit switched and packet switched services.
On the other hand, data services and telephone services can easily be
integrated on the copper lines connecting the user to the network by
simple multiplexing methods. The ISDN signaling protocol and the
multiplexing on the subscriber line also allows the ISDN user to connect
to several other users at the same time.
The two most important advantages of ISDN are that new
subscriber lines need not be deployed for data transmission and that
several simultaneous communication sessions can be established on
each subscriber line. However, ADSL and related subscriber line
technologies provide similar and simpler solutions to provide
broadband data combined with telephony over subscriber lines. ISDN
are also nicknamed as “Innovations Subscribers Don’t Need” or “It Still
Does Nothing” since the technology does not offer much more than was
already available on the PSTN and dedicated data networks.
Asynchronous Transfer Mode (ATM) was developed for supporting
both telephony and broadband data over a common packet switched
high-speed optical network (140 megabits per second). The
development of ATM took place during the 1990s coinciding with the
commercialization of the Internet. Data and speech samples are sent in
small data packets containing only 48 bytes of information. Short
packets (or cells as they are called) sent at high speed makes it easy to
avoid that the voice and video signals are destroyed by timing
fluctuations. Because of the particular routing algorithm chosen for
packet transfer, all data is sent in a connection-oriented mode; that is, a
complete connection must be established between the users before
data is sent. This applies independently of the amount of data being
sent—even the transfer of less than 48 bytes fitting into a single cell
requires the establishment and release of a connection.
At the time when the ATM specification had been finalized, the
Internet had been commercialized and had already replaced most of
the X.25 networks. Internet is a simple and cheap network, and,
initially, users were not urging for integrating voice and data
communication. Because of the popularity of the Internet and the rapid
increase in mobile communication, ATM was never realized in large
scale, and the ATM Forum promoting the technology became the
Broadband Forum in 2005 concerned with broadband network
technologies in general.

3.3 Service Integration over the Internet

The success of providing voice communications over the Internet is
related to the development of voice coding standards referred to as
voice over IP (VoIP). Development of these standards begun in 1995 by
Microsoft and Intel and in 1996 by ITU. During the early 2000s, VoIP
providers proliferated. The largest application of VoIP two decades
later is over 4G and 5G mobile networks. 4G and 5G mobile networks
offer only packet switched Internet connections for all services, and
circuit switched applications are no longer offered in the latest
generations of mobile networks.
It also turns out that the Internet is an excellent network for
providing many different types of services. The TCP/IP protocol stack
enables the offering of different services with different requirements to
quality-of-service. The flexible addressing scheme of the Internet also
enables worldwide transfer of data. The open nature of the Internet
enables a wide range of services that use the Internet as a
communication platform, including web browsing, voice and video
communication, messaging, gaming, banking, and information retrieval.
Hence, the Internet is technical capable of offering most—if not all—
services that have previously been transported over different dedicated
networks.
In 2021, it is evident that the traditional telecommunications
infrastructure is soon to be replaced by only two networks: a cellular
mobile Internet and a fixed Internet, both offering the same services.
The difference between them is the way in which the users are
connected to the network. In the cellular mobile Internet, the users are
connected via 4G or 5G land mobile networks or via low Earth orbit
(LEO) satellite systems (see the example in ► Sect. 3.4). In the fixed
Internet, the users are connected via cables or optical fibers. The major
difference between these networks is the bandwidth: it is simpler to
offer wideband data via optical fibers than over wireless connections. ◘
Figure 3.1 illustrates how this convergence has gradually taken place
during the last 30 years. The conversion process is illustrated by the
arrows from the fixed telephone service and broadcast service to the
mobile and fixed Internet services. The arrows are plotted against
technological evolutions either triggering or enhancing the process.
Fig. 3.1 The evolution from many dedicated networks to a single network—the Internet
constituting a fixed and a mobile part. Over time, services previously only available on
dedicated networks move over to the Internet. Finally—when the Internet contains all service
functionality available on dedicated networks—the dedicated network is shut down. (Authors’
own figure)

Since GSM was put into operation in 1991, the fixed-telephone

service has gradually been replaced by the mobile telephone service.
While the number of subscriptions of mobile cellular services has
tripled between 2005 and 2018 globally, the number of fixed-telephone
subscriptions is reduced by about 30% in the same period (see ◘ Fig.
3.2 (The World Bank Open Data. ► https://​data.​worldbank.​org/​)). The
situation in the Nordic countries is even more dramatic: the number of
fixed-telephone subscriptions is halved from 2013 to 2018, giving a
halftime of only 5 years (Telecommunications services in the Nordic
and Baltic countries in 2018, 2019). The smartphone market in the
Nordic countries has been saturated during the same period, indicating
that most individuals in these countries are now only using their
smartphones as their only telephone services. The same smartphone is
also most often used for both private and business purposes, also
reducing the need for the fixed-telephone service in industry and public
and private organizations. Part of the fixed telephone traffic has also
been replaced by voice over IP (VoIP) and related technologies (Skype,
Google Hangouts, and WhatsApp). However, this replacement of the
fixed telephone service was slow until 4G was put into operation
around 2010. This triggered the rapid reduction in fixed telephones we
see today. By the end of 2020, the telephone network was terminated in
Norway, and the service has been replaced by 4G and 5G mobile
networks.

Fig. 3.2 Evolution of fixed telephony and mobile cellular subscriptions. (Authors’ own figure,
data extracted from (The World Bank Open Data. ► https://​data.​worldbank.​org/​))

The mobile Internet service is the component of the general mobile

service carrying packet data over IP. The service was introduced when
the GPRS (General Packet Radio Service) functionality was added to the
GSM system in 2000. GPRS was developed further in the 3G standard,
and until 4G was released, the cellular mobile communications
networks were designed as dual networks, where one part offered
circuit switched telephone services and the other part offered packet
switched data services over a common radio interface. 4G, opened in
the early 2010s, is a pure Internet system only offering packet
switching based on IP version 6 technologies called IP Multimedia
System (IMS) and VoLTE. VoLTE (Voice over Long-Term Evolution) is a
more bandwidth efficient and flexible technology than the ordinary
VoIP technologies used on the Internet. The general international trend
is (as in Norway) that 4G and 5G mobile network technologies will
replace the telephone network. Interworking units between the mobile
Internet and the PSTN will connect regions where the PSTN still exists.
Most audio broadcast and television services are now also available
as real-time streaming services over the Internet. This is often referred
to as webcasting. Webcasting does not require a separate network of
radio transmitters, cables, or satellites to reach the audience. It only
requires broadband access to the internet, and shows may be produced
using cameras on smartphone or personal computers or standalone
webcams. The technology is cheap, and anyone may afford to produce
real-time video programs and send them over the internet—or become
a webcaster. However, there may be certain restrictions associated with
webcasting; for example, if the program contains copyrighted material
(e.g., music), the producer may have a license from the authorities to
broadcast it as a commercial product. Examples of webcasting are real-
time audiovisual dissemination of lectures, meetings, concerts,
weddings (nicknamed wedcast), and funeral ceremonies. Certain
webcasts may be viewed for free, while some are paid for per view, and
still others require prepaid subscription fees.
One type of broadcasting service is the podcast. The term was
originally used for radio programs made available as audio files that
could be downloaded to iPods. Podcast is now more generally used for
episodic series of audio or video shows that can be downloaded to MP3
players, other media players, smartphones, or computers as webpages.
Some podcasters offer subscription services where the episodes are
downloaded automatically to the user equipment.
Even though most television and radio broadcasts are available on
the Internet as webcasts and podcasts, there will still be a need for
broadcasting services over cables, satellites, and radio transmitters for
a long time to come.

3.4 Toward Next-Generation Network

We have just seen that the fixed telephone network is about to
disappear and be replaced by “all-IP” high-speed core and broadband
access networks supporting a mixture of simultaneous voice,
messaging, video, broadcast, and data services. This network is then the
first network satisfying the definition of the next-generation network
(NGN). In 2004, Study Group 13 of ITU-T defined NGN as follows (The
quote paragraph is found on the link: ► https://​www.​itu.​int/​ITU-T/​
studygroups/​com13/​ngn2004/​working_​definition.​html):
“A next-generation network (NGN) is a packet-based network which
can provide services including Telecommunication Services and is able
to make use of multiple broadband, Quality of Service-enabled
transport technologies and in which service-related functions are
independent from underlying transport-related technologies. It offers
unrestricted access by users to different service providers. It supports
generalized mobility which will allow consistent and ubiquitous
provision of services to users.”
The evolution toward the NGN has taken a long time but has
recently gained speed mainly because of the rapid rollout of 4G
networks in many countries. The reasons for the slow start were both
technological and commercial. Technological reasons were, for
example, related to low data rates in mobile systems, insufficient
capacity in transport networks resulting in large fluctuations in the
delay of data packets making them unsuitable for broadband streaming
services, slow computer hardware, and simply because it takes time to
develop and implement new technologies. The most important
commercial reason is that it is expensive to build a new network.
One approach toward new network configurations is the use of low
Earth orbit satellites. Some of these projects are reviewed in the case
study.

IP in the Sky
The feasibility of offering satellite services directly to the users of
maritime and aeronautical applications using geostationary
satellites was confirmed by several studies in the late 1960s. This led
to the establishment of the Inmarsat organization in 1981, primarily
offering satellite communications to ships and later also to aircraft.
These systems are using geostationary satellites at a height of
36,000 kilometers above the Equator. These systems require big and
expensive satellites and are not feasible for cheap broadband
communications with handheld terminals. For this purpose, several
systems using lightweight satellites in low Earth orbits (LEO) (less
than 2000 kilometers above the Earth) have been put into operation
or are planned. For more details, see the homepages for the projects
listed below.
Iridium is a system consisting of 66 satellites in polar orbits at a
height of 780 kilometers above the Earth. The original concept
included 77 satellites, and the system was named after the element
with atomic number 77 in the periodic table, iridium, since the
system looked like an atomic nucleus surrounded by 77 electrons.
The Iridium system was downscaled to 66 satellites. Element
number 66 is dysprosium, a less appealing name, so that the original
name was kept. Iridium commenced operation in 1998 after severe
initial financial problems, offering telephone services to handheld
terminals. Since 2018, the system also offers data communications
with data rates from 128 kilobits per second to 8 megabits per
second to fixed and mobile terminals.
Globalstar is an American company offering telephone services
and low-speed data communications (9.6 kilobits per second) to
portable telephone and data terminals. The system consists of 24
satellites at a height of 1400 kilometers in orbits with an inclination
of 52 degrees.
ORBCOMM offers Internet of Things (IoT) connections using both
cellular and satellite networks. The company has more than two
million subscribers in more than 130 countries. The satellite
network consists of 31 lightweight satellites in various low Earth
orbits offering low-speed data communications to IoT devices. The
company has offered satellite communications since 1990.
Teledesic was founded in 1990 to build a global satellite network
for broadband Internet communications. The system should consist
of 840 satellites in low Earth orbits. It was soon downscaled to 288
satellites. The project was abandoned in 2002 (Godwin, 2002). Bill
Gates was one of the founders of the company. Even though the
system was never built, it is interesting because the idea behind the
project lived on and may now be realized as OneWeb and Starlink
commence operation.
OneWeb is a UK-based company that launched the first satellites
in February 27, 2019. The constellation will eventually consist of 650
LEO satellites offering Internet services globally.
 Starlink is a satellite system established by SpaceX, a company
founded by Elon Musk. The system will consist of thousands of small
LEO satellites offering global broadband Internet access for
everyone. By the end of 2020, more than 1000 satellites have been
placed in orbit.
There are several other planned projects, for example, by
Facebook and Amazon.

3.5 Conclusions
Since the early 1970s, the telecommunications industry has been
striving to specify techniques by which the telephone services and the
data services can be supported by a single network. This led first to the
development of the ISDN and thereafter to ATM. Both attempts failed:
ISDN since, except for the subscriber line, it did not really integrate the
PSTN and the data network and ATM because the technology was too
late and was replaced by the much cheaper and more flexible Internet
just when it was ready for implementation.
All types of services can be adapted to the Internet technology. In
this respect, the Internet is a true integrated network. However, it took
several years before the Internet became an important carrier of voice
services so that the service integration process could really commence.
This had to wait until the 4G mobile network was put into operation in
2010.
GSM was put into operation in 1991 offering mobile telephone
services. The foremost advantage of mobile communication is
flexibility: mobile networks allow people to make and receive
telephone calls and other services anytime and anywhere. During the
next two decades, the mobile networks were developed into a mobile
Internet supporting all services—voice, data, and broadcast—by a
single technology (4G). This evolution finally resulted in a fully
integrated network, and the convergence of all services could
eventually commence. In 2020, there is technically no need for a
separate telephone network. In Norway, the telephone network was
terminated in 2020.
However, there may still be a need for dedicated networks because
the Internet may not be reliable enough to support critical
infrastructures such as distress services, energy production, and
finance.

Questions
1.
What are the major benefits and challenges of a fully converged  
ICT infrastructure?
2.
How will the cost of operating and managing ICT infrastructures  
change due to convergence?
3.
Is it possible for the Internet to accommodate all kinds of  
services offered on current dedicated networks?
4.
How will Starlink alter the competition in the  
telecommunications market?

Answers
1.
The major benefit is that there will be fewer networks to build,  
maintain, and operate. The network will offer many new
services and combine them in new ways that will support new
applications. The major challenge is competition since it will be
easier to build new networks.
2.
Managing and operation of the network will probably be  
cheaper and more efficient.
3.
The Internet offers best-effort services. This makes it difficult to  
guaranty quality-of-service objectives related to resilience,
latency, timing accuracy and jitter, loss of information, and
priority. Therefore, it may not be possible to support certain
applications (e.g., distress operations) in a proper way, hence
requiring dedicated networks.
4. Starlink will, because of global coverage, compete with all other
telecommunications operators worldwide. They will also offer  
communications to areas which are not accessible to other
t l i ti t
 telecommunications operators.

Exercise
List some of the new benefits 5G systems will offer. Search the
Internet (e.g., Wikipedia) to find out more about 5G systems.

Answers
This is some of the benefits:
– Offer wide-band wireless services in the range of 400 megabits
per second to 2 or more gigabits per second (depending on cell
size and frequencies).
– Better utilization of the frequency spectrum (less overhead and
more efficient access technologies).
– Better mix of radio cell sizes (from a few meters to several
kilometers).
– Offer edge computing, thereby reducing the latency for certain
applications, e.g., for applications on the Internet of Things. Edge
computing means that software and storage are brought close to
the user, e.g., located in the base station.
– New methods for orchestrating services consisting of multiple
applications (network slicing).
– Capable of connecting millions of devices to the same base station
(e.g., sensor networks).
– Convergence of cellular mobile networks and Wi-Fi and other
local area network technologies.

References
Godwin, R.. (2002, October 3). Teledesic backs away from satellite. ZD Net.

Telecommunications services in the Nordic and Baltic countries in 2018. (2019, December 14).
Report received from NKOM (Norwegian Communications Authority).
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_4

4. Digital Economy Ecosystem

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Ecosystem metaphor – Layered Internet – OTT services –

Transient states

Learning Objectives
After completing this chapter, you should be able to:
– Identify both direct and indirect the stakeholders in the
ecosystem of a company in the digital market.
– Analyze the impact each stakeholder has on the market
performance.
– Identify the interactions, critical relationships, and
dependencies between the stakeholders.

4.1 Ecosystem Metaphor

The concept of business ecosystems was first proposed by James F.
Moore in 1993 in a Harvard Business Review article (Moore, 1993):
“To extend a systematic approach to strategy, I suggest that a
company be viewed not as a member of a single industry but as part of
a business ecosystem that crosses a variety of industries. In a business
ecosystem, companies coevolve capabilities around a new innovation:
they work cooperatively and competitively to support new products,
satisfy customer needs, and eventually incorporate the next round of
innovations.”
Since then, the concept has been used regularly as a tool to analyze
business strategies; in particular, the complex businesses arising in the
digital economy.
The concept is a biological metaphor since many of the new digital
businesses are imbedded in a complex community of other cooperating
and competing businesses and customers much like the coevolving
ecosystems in biology. Removing one stakeholder from this community
may have severe repercussion on the businesses of the remaining
stakeholders in the same way as the removal of one species in a
biological ecosystem may sometimes alter the whole ecosystem system
in a negative way or even destroy it. ► Box 4.1 illustrates the
vulnerability of a biological ecosystem where the removal of a single
species destroys the whole ecosystem. Industry may be equally
vulnerable to apparently small alterations in the business environment,
for example, change in regulations, customer habits, discovery of new
raw materials, and better production methods.
The biological metaphor is a fruitful analogy since the complexity
and unpredictability of the evolution of the information technology is a
result of visible and hidden as well as planned and unplanned
interactions within a business ecosystem consisting of entrepreneurs,
researchers and developers, software and hardware platform
providers, system integrators, network providers, government
authorities, and communities of individuals participating directly or
indirectly in the evolution (Muegge, 2013). Examples of systems
resulting from such complex and unpredictable interactions are the
Internet itself and the smartphone technology consisting of an intricate
hierarchy of devices, services, and applications. Peer-production,
crowdsourcing, and the use of free and open software make the
ecosystem even more multifaceted and unpredictably complex (see ►
Chap. 7).
The digital economy depends intimately on the information and
communication technology (ICT) and consists of elements such as the
Internet, smartphones, data storage, and processing. Put differently,
digital services require ICT to manage and propagate its value
proposition to the consumers. The interplay between digital services,
information and communication technologies, providers of digital
services, competitors, digital marketplaces, consumers, and society at
large is the core feature shaping the digital economy. This interplay is
called the digital economy ecosystem
.

Definition 4.1 The digital economy

ecosystem
describes the
relations and dependencies between digital services, ICT
infrastructures, digital markets, and authorities in a socioeconomic
context.

Box 4.1 Illustration of the Vulnerability of a Biological

Ecosystem
The vulnerability of a biological ecosystem is illustrated by a small
but well-documented ecosystem in the salt marches in Southern
California (Lafferty & Morris, 1996). The ecosystem consists
primarily of four species: the parasite Euhaplorchis californiensis,
horn snails, killifish, and seabirds. The horn snail and the killifish are
intermediate hosts for the parasite. The bird is the final host in which
the parasite matures sexually and produces eggs that are spread by
the bird’s feces that are eaten by the snails. This completes the life
cycle of the parasite as shown in ◘ Fig.  4.1.

Fig. 4.1 Biological ecosystem. (Authors’ own figure)

 The parasite alters the behavior of the intermediate hosts: the
snails search for water where they are eaten by the killifish; the
infected fish is no longer afraid of birds and becomes an easy prey. In
one area, the authorities decided to remove the parasite to make the
beach healthier. This caused a complete collapse of the ecosystem:
the snails no longer searched for water, the killifish disappeared
since there was no more easily accessible food close to the shore, and
so did the birds because easy prey was no longer available.

4.2 Ecosystem Components

The only way to appreciate the complexity of the business ecosystem of
digital services is to identify the basic technologies and building blocks
it contains.
◘ Figure 4.2 illustrates the digital economy ecosystem and some
basic technologies and stakeholders on which it depends. The global
ICT infrastructure—owned by the Internet service providers (ISPs) and
the network providers (NP)—is the carrier of digital services, including
applications and content. This ICT infrastructure consists of
interconnected networks of networks including the Internet, mobile
networks, wireless networks, local area networks, fiber networks, and
satellite networks. It also includes storage of data and computing
facilities. The ICT infrastructure is accessed by various types of user
equipment—devices—including mobile phones, smartphones, PCs,
laptops, and other terminals. Such user equipment is manufactured and
offered by device providers (DP). Application service providers (ASPs)
offer applications to the consumers by trading them in the digital
marketplace. Applications run on devices and may feature content
offered by content providers (CP). An example of this is YouTube—an
ASP—offering video content to users, in which the producers of content
are CPs. Consumers and providers interact with authorities adapting
legal frameworks and satisfying societal demands. To understand this
complex environment, it is necessary to investigate how the different
elements of the digital ecosystem interact with one another.
Fig. 4.2 The digital economy ecosystem. (Authors’ own figure)

A digital service usually depends on other digital services to provide

value to consumers. Existing digital services are also the basis for the
development of new digital services. For example, the business of
Amazon depends on digital payment services to offer e-commerce, and
Spotify depends on access to digitized music to provide streaming
services to consumers. These services may, in turn, depend on various
functionalities offered by several ICT providers and other providers of
hardware and software support functions. Together, these
dependencies make up the digital economy ecosystem for the
particular business. It may be impossible to identify all dependencies
merely because of complexity and transient relationships that may
arise and vanish during the production of the good. The complexity of
the ecosystem is exemplified in ◘ Fig. 4.3 where a digital Service A
depends on two other digital services (B and C) as well as two
technologies (cloud storage and wireless access). Furthermore, Service
E depends on Service A, while Service D and Service A depend on each
other. ◘ Figure 4.3 also suggests a simple method by which the
ecosystem can be modeled to analyze the impact each dependency has
on the primary business. This analysis may also uncover critical
dependencies, unexpected vulnerabilities, structural weaknesses, and
other problems.

Fig. 4.3 Dependencies in the digital economy ecosystem. (Authors’ own figure)

The complexities and dependencies in the digital economy

ecosystem result in a market that is never the same from one day to the
next. Changes such as new stakeholders entering the market,
development of new technologies, variations in consumer adoption,
altered competition arenas, new legislation and market regulations,
and fluctuating dependencies among the stakeholders, all contribute to
a market state that is more transient in nature than stationary. ◘ Figure
4.4 illustrates the difference between transient and stationary states.
The transient view of the digital economy is consistent with the
teachings in complexity economics (Durlauf, 1998)
. In complexity
economics, the basic argument is that the economy is ever-changing
and will never reach a stationary state. The reasons for this are rooted
in the digital economy ecosystem and the rate of technological
innovation
.

Fig. 4.4 Transient and stationary states. (Authors’ own figure)

For a stationary market to exist, most of the market aspects

mentioned above must be stable and unchanging. In the digital
economy, new technologies are, in contrast to stationary markets,
developed and adopted at a rapid rate. A good example is digital mobile
communication which has changed from supporting simple telephone
and messaging services (GSM in 1990) to offer a mix of telephone
services, messaging services, data services, broadband streaming
services, and IoT (5G in 2020) by adding new service capabilities,
upgrading the network architecture, and implementing new access
technologies every few years.
 New technologies often lead to new business models, thereby
causing changes in the digital economy ecosystem. In mobile
communications, this has been an evolution from simple services
offered by a single telephone services provider to services offered by an
amalgamation of almost all kinds of application service providers
(ASPs). All this contributes to the transient nature of the digital
economy. ► Chapter 18 presents quantitative models for analyzing the
transient nature of the digital economy.
The rest of this chapter contains more details concerning the most
important components of the digital ecosystem: the Internet, computer
platforms, applications, consumers, and authorities.

4.3 The Layered Internet as Ecosystem

Component
One of the most important events in the evolution of digital markets
was the commercialization of the World Wide Web (WWW) in 1993.
Before 1993, network operation and service provision were integrated
industries, and consumers could not freely choose network access
independently of service. The WWW led to a restructuring of the ICT
business in which network operation and service provision became
independent industries. This was possible because of the layered
technological architecture of the Internet.
◘ Figure 4.5 shows the layered Internet model in which the Internet
is divided into three parallel planes: networks, user equipment, and
applications. The three planes correspond to three independent
business areas: (1) networks; (2) the development, production, and
sale of user equipment; and (3) the provision of services, information,
system management, and remote sensing and control (Audestad, 2007).
Fig. 4.5 The layered Internet model. (Authors’ own figure)

The networks plane shows the physical communication network as a

graph in which communication links (physical links) interconnect the
routers using the Internet protocol (IP). This graph then symbolizes the
physical network supporting the businesses of the network providers
(NPs) and the Internet service providers (ISPs). This is, strictly
speaking, what is defined as the Internet.
The user equipment plane consists of a graph in which the hosts
(e.g., computers, servers, and databases) are nodes, and a connection
between two hosts indicates that the hosts are taking part in a common
computation. This is the transport layer, and TCP/UDP
/SCTP1
connections exist between hosts. The user equipment plane is the
business domain of the device providers (DPs).
The application plane is a graph in which software objects
(applications) are nodes. A connection between two nodes (often
referred to as virtual connection) indicates that the corresponding
software objects take part in a common computation. A smartphone
and downloaded apps are nodes in such a graph. This graph is dynamic
and may alter configurations in milliseconds. This may then include
complex configurations such as cloud computing, software-defined
networking (SDN), virtualization (NFV), and service-oriented
architecture (SOA). The application plane is the business arena of the
application service providers (ASPs) and the content providers (CPs).
Since software runs on computers and computers are connected to
the Internet, there also exist vertical interactions (inter-layer
connections) between the layers as shown by dotted lines. Note that the
structures of the graphs on each layer are independent of one another
except that there must be a path in the network plane connecting two
hosts in order for a TCP/UDP/SCTP connection to exist between them.
A TCP/UDP/SCTP link must exist between two software objects
running on different hosts to take part in a common computation.
Note also that competition takes place between stakeholders on the
same layer of this model. There is no competition between operators on
different layers. The user equipment layer can then be regarded as the
demarcation line between the business segments of the ISPs in the
network plane and the ASPs in the application plane. The important
point is that the ISP cannot perceive what kind of services is carried by
the network since this is obscured by the protocol on the transport
layer (TCP
/UDP
/SCTP
) itself and the Transport Layer Security (TLS)
encryption protocol in particular. This has a severe impact on pricing
regimes available to the ISP: since the ISP cannot identify the type of
service, the ISP cannot levy differentiated charges. The ISP is then
forced to apply service-independent charging. This separation of
services and networks is also one of the core premises of net neutrality
(see ► Chap. 21)
—the ISP cannot discriminate between different uses
of the network and, therefore, cannot levy charges reflecting the value
of the service offered by the ASP.
 The businesses of the ISP and the ASP are independent in the sense
just described. The businesses of the ISP and the ASP are also
complementary in the sense that each needs the other to conduct its
own business. The ASP cannot offer services without the assistance of
ISPs, and the ISP does not create value for the consumers beyond that
of transferring bits and providing basic communication services. Hence,
the ISP needs the ASP to make the network attractive and useful for
consumers, and the ASP need the ISP to deliver the services to the
consumers.
One particular consequence of this decoupling of ASP and ISP
functionality is over-the-top (OTT) media services in which the ASP
offers services that traditionally were supported by dedicated networks
—such as the telephone network and cable television networks—
without any involvement by the ISP other than sending the information
in unspecified IP packets to the consumers. See ► Box 4.2 for Skype as
a typical over-the-top telephone service. Note that voice over IP (VoIP)
may not always be an over-the-top service since several VoIP
applications require particular protocol support from the ISP, for
example, using the Session Initiation Protocol (SIP) for establishing and
deleting the call and supporting quality of service management. In this
case, the ISP can distinguish VoIP from other Internet services and levy
charges based on this knowledge. The telephone service in 4G and 5G
mobile systems is a VoIP service of this type, but competition from
Skype has limited the opportunities of mobile operators to levy
differentiated telephone charges based on call duration and distance.

Box 4.2 Skype and Over-the-Top Services

Skype is a digital service providing chat, voice, and video
communication for Internet users. Skype was launched in 2003 as a
desktop service. It was bought by eBay in 2005 for $2.6 billion and by
Microsoft in 2011 for $8.5 billion. In 2013, Skype had almost 40% of
the international voice and video market (Gara, 2014).
Skype is a contributor to the ongoing convergence of digital
services along with other providers of voice-over-IP services such as
Google Hangouts and appear.in. These are examples of over-the-top
(OTT
) services. They provide free voice, video, and messaging over
the Internet, thereby competing with the traditional telephone and
video services. The introduction of Skype and other voice-over-IP
services has had a significant impact on the digital economy in which
the main outcome is a dramatic decrease in prices for telephone calls
both nationally and internationally. Mobile network operators offer
free telephone services in their subscription schemes to meet the
competition from Skype and other similar OTT services.
At the time Skype was launched, the users of traditional
telephone services had to pay for the volume, duration, and distance
of the calls they made. Skype undermined these payment methods by
offering free calls between Skype users anywhere in the world. It is
no wonder why people abandoned the expensive traditional
telephony services and adopted Skype, especially as a replacement
for long-distance calls. Skype is a typical value network, mediating
communication between users. The company has high fixed costs
(e.g., software development and infrastructure costs) and low or
close to zero marginal costs. The value of Skype corresponds to the
size of the network; that is, the number of users.
All web services, including Skype, use either http or https
(encrypted http access) as carrier protocols. It is, therefore, not
possible to distinguish between different services from the type of
protocol only. Moreover, the web services include most of the
services on the Internet. The internet service provider cannot levy
time and volume-dependent charges for web services (e.g., a fixed
price per byte) because then some services would be too expensive
(e.g., video streaming services), while other services would cost
almost nothing (e.g., email). This is net neutrality in practice: the ISP
cannot discriminate data on the Internet based on type of content.
For this reason, the ISP can only take fixed access charges, often
based on the amount of data that can be downloaded per month as
prescribed in the subscription contract plus additional charges if the
user downloads more than the contractually agreed amount of data.
On the other hand, the ASP may levy differentiated fees for apps,
films, and music.
OTT
services in general drives the convergence of digital services
by providing the traditional teleservices at a lower cost over the
Internet.
4.4 Computer Infrastructure and Platforms as
Ecosystem Components
The software modules required to produce a digital service run on
computer platforms. The platform is made up of the equipment in the
equipment plane of ◘ Fig. 4.5 and is the mediator between the network
infrastructure and the application software. The platform may consist
of a single computer or be a distributed system of interconnected
computers, possibly owned by different organizations or people. The
overall platform architecture may, therefore, be very complex and
involve several stakeholders who must cooperate to provide the service
offered to the consumers. Google maps showing the GPS location of the
user is an example of a service requiring at least two independent
providers of partial information.
The most common platform is cloud computing offering services
ranging from pure infrastructure support to complete software
packages and any combination thereof. In the cloud computing
business model, the provider of the digital service or good does not
own all the production facilities but rent some of them from other
stakeholders (the cloud providers). The service provider then reduces
the need for investments in ICT infrastructure and can focus entirely on
the core business. Other advantages are that the provider can reduce
the time to develop the product considerably, meet challenges
associated with fluctuating market size, eliminate the need for expert
knowledge of processing and storage technologies, and reduce the need
for system management and maintenance.
The key concept in cloud computing is “Everything as a Service”
with acronyms EaaS or XaaS. There are three main categories of
services:

Infrastructure as a Service (IaaS)

In this case, the client rent access to the IT infrastructure of the cloud
provider (computers, databases, servers, communication interfaces,
and network). The IaaS provider offers a virtual machine architecture
supporting a number of simultaneous and functionally separated
clients. Each client will then view the cloud as an individual computer
where the client may develop and run arbitrary software, also including
own operating systems and security protection if required. The client
has no control over the underlying infrastructure but may have control
over operating systems, storage, use of software libraries, and security
settings.

A related concept is network virtualization where the user is not aware

of the fact that the processing activities are distributed over several
computers. This implies that the user perceives the system as a single
computer over which the user has exclusive control.

Platform as a Service (PaaS)

In this case, the cloud provider offers a software development
environment for the clients. The PaaS platform may contain
programming tool kits, software libraries, as well as operating systems,
compilers, and databases. App Engine is a PaaS platform offered by
Google for development of web applications.

Software as a Service (SaaS)

In this case, users are given licenses to use application software and
database functions. All processing is then managed by the cloud
provider. Examples of SaaS are access to office software, maps,
Geographic Positioning System (GPS), Geographic Information System
(GIS), human resource management, gamification, collaboration,
learning, and many other applications.

Grid computing is the interconnection of many heterogeneous

computers to perform particularly large computational tasks. The grid
is different from the cloud in the sense that the computers involved are
lightly connected, and each of them performs a dedicated and invariant
task. The grid may then be viewed as a vast distributed supercomputer.
The number of active computers in the grid involved in the same task at
any instant is also varying since the computers are usually used by the
grid only when they are not busy with computations initiated by the
owner of the computer. Typically, the grid consists of computer facilities
at universities and research institutions.
Tasks using grid computing are computationally hard problems
such as studying protein folding processes, simulating climate models,
searching for large primes, and analyzing particle collisions at CERN.
There are several active international grids.

4.5 Applications and Content as Ecosystem

Component
The number of applications available to customers is enormous, in
particular, after the smartphone entered the market in 2012. The
number of available apps in Google Play Store and Apple’s App Store
were about 3.0 million and 3.4 million, respectively, by the end of 2020
(Number of available apps in the Apple App Store from December 2020.
Statista. December 10, 2020; Number of available applications in the
Google Play Store from December 2009 to September 2020. Statista,
December 10, 2020).
Some digital products are produced and sold by a single ASP only
involving the ISPs for transporting the information to the customers.
The biggest companies in the digital business, e.g., Facebook, Twitter,
YouTube, and Google, are conglomerates consisting of several
subsidiaries providing most of the content associated with the
company. They also run their own computer platform, and some of
them even offer ISP services and cloud computing.
Services may also be complex, involving not only an ASP but also
several content providers and platform providers. One example is
newspapers. A newspaper may buy or receive free information from
other newspapers or magazines and engage freelance journalists or
photographers. The newspaper is also member of a web of text and
video content providers, news agencies, and stock photography
agencies.

4.6 Consumers as Ecosystem Component

The consumers of digital service providers may be companies,
organizations, or individuals. Examples of companies acting as
consumers are electric power providers receiving automatic
consumption measurements form censor platforms and newspapers
buying pictures, news, and other topics from content providers.
 The consumers in digital markets play an essential role in the digital
economy not only as consumers of digital goods but also as a source of
feedback to designers and providers of the goods that may be used to
improve the product. The importance of market feedback is
emphasized in ► Chap. 9. Closely related to market adoption is the
effort by the authorities to regulate both the digital product and the
marketing of it.
Consumer adoption (or user adoption) of information and
communication technologies and digital services is a key element to
apprehend the creation of dependencies in the digital economy
ecosystem. Consumer adoption is a measure of how and at what rate
consumers are adopting a specific digital service. Some aspects
concerning consumer adoption rate can be summarized as follows:
– Obviously, the technology or digital service must be adopted by some
initial consumers to have any impact on the market evolution. If
there are too few initial consumers, then evolution may stagnate, and
the technology or service may disappear from the market. Examples
of this are videophones only enhancing ordinary telephony by
showing videos of the speakers while talking and the Teletex service
for transmitting documents over telephone networks. The
videophone service required expensive user terminals and offered
the users only small advantages compared to the simple and cheap
telephone service. Moreover, the videophone service was subject to
strong network effects: the first videophone has zero value since
there are no other communication partners. The timing of the
introduction of Teletex was bad, and the service was almost
immediately substituted by the free-of-charge e-mail service over
Internet.
– Early market adoption is particularly important if the market growth
depends on network effects (see ► Chap. 18)
. Social services, such
as Facebook, are examples of services that are subject to strong
network effects having difficulty attracting early users.
– The adoption of a technology or digital service may depend on other
technologies or digital services to be widespread and fully adopted
by consumers. For instance, Uber would have never become a
success without the high adoption rate of smartphones and wireless
Internet access. The video-on-demand service requires broadband
 access by users and depends on optical network technologies for
sufficient traffic capacity in the network and for access to the
database.
– When a technology or digital service starts to attract consumers and
the use of it increases, it may trigger the evolution of new
technologies or services. The two most obvious examples are
Internet and mobile networks prompting the entire evolution of
digital services. A less obvious example is touchscreens for mobile
terminals stimulating the development of apps.
– New dependencies and entirely new stakeholders may appear as the
digital service evolves and gets adopted by even more users. Offline
use of credit cards prompted new operators to offer automatic card
reader facilities for direct payment in shops and hotels. This has
further stimulated the evolution of card readers in petrol pumps,
parking meters, public toilet locks, and payment automata in public
transport. Apps on mobile phones are now taking over several of the
applications previously supported by credit cards.
Hence, analyzing the targeted technology or digital service in
isolation will not predict how attractive it will be—several other
services and functionalities needed for the adoption of a technology or
a digital service must also be considered. All these interdependencies
make up the ecosystem of the digital service. A quantitative consumer
adoption model based on the Bass equation is presented in ► Sect. 18.​
2.

4.7 Authorities as Ecosystem Components

The digital economy ecosystem also consists of social and legal aspects
concerning conditions for marketing and use of digital services,
including how technologies and digital services are regulated by the
government. Legal regulations of digital services may not only influence
the content of these services but also how they are adopted by the
consumers. Such legal regulations may also have an impact on the
visibility, development, and use of related digital services. Examples of
the impact of legal aspects on the digital economy are the regulations of
cryptocurrencies and sharing services such as Uber (Court of Justice of
the European Union, 2017) and Airbnb (Tuohey, 2018)
. The decision
by the European Court of Justice (ECJ) to classify Uber as a transport
company has significant bearing on the business operations of Uber.
The authorities also follow the businesses conducts of the large
corporations in the digital economy closely to uncover, for example, tax
evasion, unethical or illegal exploitation of personal information
gathered about consumers and users, infringements or circumventions
of competition laws and regulations, building of lock-in barriers to
create monopoly advantages, unsustainable environmental footprint
and use of energy, exploitation of workforce, and negligence regarding
child work. Most of the large corporations in the digital economy have
been accused of misconduct in one or several of these areas. However, it
has been difficult to alter the behavior of these corporations, though
criticisms from authorities and the public, and sometimes legal
proceedings, have forced the corporations to alter their business
conduct in some cases.
The role of the authorities as watchdogs in the digital business
sector is, therefore, important to avoid concentration of too much
power on just a few stakeholders in the digital economy.

4.8 Conclusions
The ecosystem metaphor is an important tool for strategic planning in
the digital economy. The ecosystem analysis reveals, among others,
stakeholders that may influence the business performance directly or
indirectly, the type and strength of the interactions with these
stakeholders, and possible vulnerabilities and weaknesses with the
chosen business model.
The ecosystem does not only consist of cooperating and competing
companies but also less obvious stakeholders such as consumers and
authorities. To estimate the performance of the company or the mere
survivability of it in the marketplace, public opinion and government
regulations must also be considered.
Some ecosystem aspects are related to the technology such as the
underlying Internet and access technology, own computer
infrastructure and computer infrastructure owned by other parties
(e.g., cloud services), and the development environment for apps and
other software and hardware components.
 The aspects outlined in this chapter are applied in ► Chap. 19
where several case studies are presented.

Questions
1.
Which information and communication technologies (ICTs) does  
Dropbox depend on?
2.
What digital services depend on Dropbox?  
3.
What impact has the authorities on the business operations of  
Google?

Answers
1.
Internet, smartphones, personal computers, cheap mass storage  
of data, cryptography, digital payment solutions, and mobile
apps.
2.
Cloud storage, file sharing, file synchronization, and project  
collaboration.
3.
Google’s business operations have been under investigation by  
both the US authorities and the EU. Authorities in the USA have
filed a lawsuit against Google in 2020. The EU has fined Google
several times in the past decade for a total of €8 billion.

Exercises
1.
Describe the ecosystem for Apples App store.  
2.
Describe the ecosystem of Uber.  
Answers
 1. Apple, app developers using the app development platform of
Apple, third parties required for management of the particular  
app (e.g., providers of supplementary content, payment systems,
cloud computing), mobile network operators and the
technological platform they offer, other network providers,
smartphone manufacturers (e.g., operating system), standards
organizations (e.g., 3GPP) specifying the basic technology
supporting apps and other content, regulatory authorities (e.g.,
violations of law and license conditions), and app users (e.g.,
usefulness of app, cost of usage).
2.
Uber depends on technologies such as the smartphone, mobile  
broadband networks (e.g., 4G and 5G), GPS, and digital payment
system, among others. In addition, Uber depends on local
authorities and the legislations in the areas they operate. Since
Uber hires drivers as a part of their business model, they will
have formal contracts with these drivers and may also be
dependent on labor unions and other local rules.

References
Audestad, J. A. (2007). Internet as a multiple graph structure: The role of the transport layer.
Information Security Technical Report, 12(1).

Court of Justice of the European Union. (2017, December 20). The service provided by Uber
connecting with non-professional drivers is covered by services in the field of transport.

Durlauf, S. N. (1998). What should policymakers know about economic complexity? The
Washington Quarterly, 21(1).

Gara, T. (2014). Skype’s incredible rise, in one image. The Wall Street Journal.

Lafferty, K. D., & Morris, A. K. (1996). Altered behavior of parasitized killifish increases
susceptibility to predation by bird final hosts. Ecology, 77(5).

Moore, J. F.. (1993, May–June). Predators and prey: A new ecology of competition. Harvard
Business Review.

Muegge, S. (2013, February). Platforms, communities, and business ecosystems: Lessons learned
about technology entrepreneurship in an interconnected world. Technology Innovation
Management Review.
Tuohey, P. (2018, January 10). Cities and state are struggling to regulate Airbnb. The Hill.

Footnotes
1 TCP = Transmission Control Protocol for connection-oriented data transfer, UDP = User
Datagram Protocol for sending of uncorrelated data packets in the connectionless mode, SCTP =
Stream Control Transmission Protocol for supporting both connection-oriented and
connectionless services in the same data transfer session

 
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_5

5. Digital Market Evolution

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Natural monopoly – De-monopolization – Market

liberalization – Resellers – VNO – MVNO

Learning Objectives
After completing this chapter, you should be able to:
– Explain the process leading to de-monopolization of
telecommunications in Europe.
– Understand the consequences de-monopolization has on the
market structure in telecommunications.
– Explain the difference between the three concepts: reseller,
virtual network operator, and mobile virtual network
operator.

5.1 Telecommunications as Natural Monopoly

The telecommunications industry has undergone an evolution in
market structure from monopoly to competition market as illustrated
in ◘ Fig. 5.1. This process is referred to as the de-monopolization of the
telecommunications market. Other often used terms are market
deregulation and market liberalization. This chapter describes how this
evolution took place in the European Economic Area (EEA).

Fig. 5.1 Evolution of the telecommunications business. (Authors’ own figure)

The evolution in EEA took place in three steps:

1.
The market for retail sales of user equipment was opened for  
competition during the period 1985–1987.
2.
Competition was introduced for mobile network operation, first in  
the UK (1982) and about 10 years later in other EEA countries
(1991).
3.
Full competition on all aspects of telephone network operation in  
Europe was introduced in 1998 (in 1996 in the USA).
Traditionally, most telecommunications operators in Europe were
state-owned monopolies. There were also privately owned telephone
and telegraph companies, but these companies had monopolistic rights
to offer telecommunications services in particular regions of the
country. The argument in favor of monopolies was that it would be
more expensive for the users if there were more than one telephone
operator in the region because of the large investments in
telecommunication infrastructures required. Moreover, the technology
used prior to the 1980s (electromechanical telephone exchanges
interconnected by coaxial cables and radio relays) had an economic
lifetime of several decades, often as much as 50 years. Therefore, it was
deemed inefficient to allow several telecommunications carriers to
build their own communication networks delivering the same set of
services. Telecommunications was then regarded as a natural
monopoly.
The state monopoly owned the network, offered the few services
supported by the network, and sold or rented out telephones, local
switchboards, data modems, and other terminal equipment. These
telecommunications carriers were called vertically integrated
monopolies. Consumers usually had one choice concerning network
provider, telecommunications service, and type of user equipment. The
governments also decided the charges the subscribers had to pay for
subscriptions and use of the services.
The situation was more complex for long-distance communication
systems such as intercontinental submarine cables and satellite
systems. Several consortia owned competing intercontinental cable
systems, and the international organization Intelsat was a competitor
to these systems offering an alternative for long-distance
interconnections over geostationary satellites. These systems were not
subject for debate in the de-monopolization discussion that followed.
During the late 1980s, it was questioned whether it would be better
to open up for full competition in telecommunications considering the
rapid evolution of digital networks and digital switching, the growing
need for computer communications, and advances in mobile network
technologies. This came at the same time as the internationalization of
the industry started in general. Many companies expanded to become
international corporations with factories in several countries. This
evolution also triggered the governments to consider opening up
national monopolies for full competition to enhance innovation and
making services and industrial products cheaper for the consumers.
De-monopolization and the belief in free markets became the zeitgeist
of the late 1980s. However, the process to transform the monopolistic
telephone operators into competitive businesses in a competitive
market took a long time because new competition laws and market
legislation had first to be put in place and enough time had to be
allowed for the monopolies to reconfigure their business models to face
a situation where they had to fight for market size and revenues.
5.2 De-monopolization of User Equipment
In the early 1980s, the first public data networks were put into
operation, and the first automatic mobile networks were up and
running. The number of different types of user equipment had
exploded, and the monopolies were too bureaucratic and too inexpert
to handle this profusion of new equipment. Responding to this, starting
from about 1985, the authorities opened the sale of user equipment for
free competition; however, the equipment had to be approved by the
telecom operator or a separate regulatory authority before the new
device could be connected to the network to ensure that the equipment
met international and national performance standards.
The number of independent retailers of various types of user
equipment grew rapidly; in particular, for sales of ordinary telephones
and mobile phones. An important offspring of the deregulation was that
the telecommunications operators no longer owned the telephone
apparatus, the data modem, or the local switchboard at the user
premises as they did before sale of user equipment was opened up for
competition. This equipment was regarded as a technical extension of
the network and, as such, an integral part of the network. After the
deregulation, the operator’s responsibility and ownership of equipment
ended at the network interface device (NID) on the wall of the house;
this technology is often referred to as “wire-to-the-wall” and, in the
optical age, “fiber-to-the-premises.” The manufacturers could now build
the data modem into, for example, computers, fax machines, and
copying machines. This simplified the use of data communications but
had little impact on the number of users of data communications until
the Internet was incorporated in the portfolio of the
telecommunications operators in the mid-1990s.

Definitions 5.1
Terms often used in the literature related to the local wire, cable, or
fiber interconnecting the subscriber and the telecommunications
network are the following:
– The local loop is the connection from the local telephone exchange
(or Internet router) to the premises of the subscriber.
 – The last mile refers to the same part of the connection.
– Network interface device (NID) is the demarcation point between
the local loop and the internal wiring at the user’s premises. The
responsibilities of the operator end at the NID.

The first regulatory authorities were established during this period to

ensure fair competition and to avoid that the telecommunications
monopolies misused their market power to hinder other retailers to
establish their independent businesses. The regulatory authorities also
issued licenses for sale of equipment and followed up that the retailers
had access to enough technical expertise for installation and
maintenance of equipment.

5.3 De-monopolization of Mobile Network

Operations
In 1981, the Nordic Mobile Telephone (NMT) had just been put into
operation in the Nordic countries. NMT was the first cellular system
offering automatic roaming and undisruptive handover of calls when
the mobile terminal moved into a new cell. Already in 1982, NMT was
about to become the preferred common European land mobile system.
British Telecom participated in this project. In 1982, Prime Minister
Margaret Thatcher and her government decided there should be full
competition on mobile communications in the UK with two
independent operators. This implied that the UK had to choose a
system other than NMT; otherwise, one of the competitors would have
too big advantage. Europe was then left with four incompatible
automatic land mobile systems: NMT in Norway, Finland, Sweden,
Denmark, Iceland, Spain, the Netherlands, and Switzerland; TACS in the
UK and Ireland; C Netz in Germany; and Radiocom 2000 in France.
This was, in fact, the major incentive for the Netherlands to suggest
in 1982 that Europe should develop a new pan-European digital mobile
system—the Global System for Mobile Communications (GSM). GSM
was originally an abbreviation for the name of the group developing the
technology—Groupe Spécial Mobile. In 1992, the GSM system was put
into operation, and EU and EFTA decided that each country should have
at least two competing land mobile networks. GSM was an ideal place
where the de-monopolization of telecommunications could start. The
countries developing the GSM standard had already agreed that the
whole telecommunications business should be de-monopolized soon.
GSM was a completely new network where all operators had to build
the network infrastructure from scratch. The new infrastructure
consists of base stations, telephone exchanges supporting entirely new
functions, and entirely new databases for subscription handling and
location management. The only advantage the telephone monopolies
had was transmission lines that could be used to interconnect the new
devices, thereby reducing the need for investments in basic
infrastructure; however, by simple regulatory requirements, all mobile
operators in the region had equal opportunities to lease such lines from
the monopoly operator for the same price as a subsidiary of the
monopoly operator.
The fixed network operators would remain monopolies offering
fixed telephone services. Hence, from 1992 onward, consumers could
choose between at least two providers of mobile telecommunications
services in Europe.
A mobile operator established in one country could now also
establish subsidiaries in other countries, thereby increasing the market
of potential subscribers and, as a result, enhancing its business
prospects and boosting its financial value. Several mobile
telecommunications companies then rapidly developed into large
international conglomerates.
One particularly amusing strategic dilemma that this situation led to
is illustrated in ► Box 5.1.

Box 5.1 Competitive Dilemma

◘ Figure 5.2 illustrates the competitive dilemma many mobile
services operators are facing, using the mobile telephony market in
Norway and Sweden as an example.
Fig. 5.2 A strategic dilemma between Telia and Telenor. The markets shares are based on
2019 figures published by NKOM (Norwegian market) and PTS (Swedish market). (Authors’
own figure)

Telia and Telenor are dominating mobile operators Sweden and

Norway, with market shares (in terms of subscribers) of about 35%
and 60%, respectively (per 2019). Telenor is a minority operator in
Sweden, and Telia is a minority operator in Norway with market
share of 19% and 33%, respectively. This leads to a situation where
the operators have to instigate different strategies in the two
countries. In Sweden, the strategy of Telia is to hinder Telenor to
capture market shares, while the strategy of Telia in Norway is to
capture as much of the market from Telenor as possible. The strategy
is, of course, the same for Telenor: gain market shares in Sweden and
protect its market shares in Norway. The two strategies are
 incompatible, and the mobile services operator must master each of
them.

5.4 De-monopolization of All

Telecommunications Operations
In 1998, the EU opened all aspects of telecommunications for full
competition. The process toward full deregulation had started already
in 1987 by the Green Paper on the Development of the Common Market
for Telecommunications Equipment and Services (Lando & The
European community’s telecommunications deregulation, 1994). One
important requirement and a first step toward full deregulation was
that the countries separated the telecommunications operator from the
telecommunications policy maker. Previously, the telecommunications
operator (or PTT as it was called in many European countries1) had
both roles. The earlier monopole operator was referred to as the
incumbent operator.

Definition 5.2
The telephone operator (or carrier) that existed before de-
monopolization is usually referred to as the incumbent operator (or
just the incumbent) to distinguish it from new operators. The
incumbent had the advantage of having built the network
infrastructure from “monopoly money” or by government subsidies.
This advantage had to be eliminated by govern regulations before
real competition could take place.

Note that in 1998, fixed-telephone services were still regarded as the

most important business in the telecommunications industry despite
that mobile phone and data services were growing rapidly. Mobile
communications had already been de-monopolized. The Internet had
existed for several years as an independent network not owned by
anyone. In 1998, Internet had just started to be included in the business
portfolio of the telecommunications operators in Europe but was still
regarded as a rather minor addition to their portfolio. Almost
unrecognizably, the Internet had started to replace the X.25 data
network as carrier for data communication services. While the network
operators could levy differentiated charges on the services offered by
the telephone network (local calls, long-distance calls, international
calls, calls to value-added services, and so on), it turned out that this
was not possible for Internet services. The revenue basis of the
telecommunications industry was about to change.
Both the Internet and the mobile phone have altered the business
landscape of telecommunications entirely. Now, about 20 years after
de-monopolization, the fixed telephone service is about to be replaced
by cellular mobile networks, and the telephone service, fixed or mobile,
is itself soon incorporated as one out of numerous data services on the
Internet using voice over IP technology. This is the convergence of
technologies and services described in ► Chap. 3.
The deregulation process took several years because the
telecommunications network was regarded as a public utility that was
best served by the old state monopolies (the incumbents). Moreover, it
was a long and difficult process to establish the rules and procedures
for regulating the market so that new entrants had a fair chance to
compete with the incumbents. As already mentioned, the GSM network
was an exception because this was a new network and the cost of
building the network was regarded to be the same for all competitors. If
the incumbent also owned a mobile network, this network had to be
commercially separated from the telephone network, and no cross
subsidizing and other value exchange was allowed between them.
Moreover, if required by the competitor, the incumbent had to provide
leased lines, at a competitive price, as feeder lines to base stations and
other equipment.
In the USA, the telecommunication market was opened for free
competition in 1996. The rules governing the competition were stated
in the Telecommunications Act of 1996. Note that the divestiture of
AT&T in 1984 was not de-monopolization of telecommunications in the
USA. AT&T was broken up into seven regional monopolies offering
telecommunications services in non-overlapping regions. The purpose
was to reduce the market power AT&T had built up over several years.
Telecommunications still existed as a monopoly business until 1996.
After 1998, anyone in the EEA could become a network operator,
service provider, or retailer of user equipment. However, the
stakeholders in this market were subject to some regulatory
restrictions related to the competition between network operators—
including virtual network operators—on price, performance, customer
care, and quality of service. These regulations included mandatory
cooperation between network operators to ensure full connectivity
between users of competing networks at reasonable prices and quality
of service and non-discrimination of application service providers
accessing the network, in particular, preventing network operators
from giving advantages to application service providers owned by
themselves.
To understand the present situation, it is important to note that the
deregulation of 1998 had to do with the telephone network only. The
driving force for the de-monopolization was the political idea that a
competitive market would be more efficient and offer lower prices than
the monopoly. This conclusion may be true for fixed and mobile
telephone operation, but the development of Internet services has
shown that this is not always true. A concern for policy makers now is
that free competition has led to the undesirable situation that several
companies in the data or Internet business have had a tremendous
increase in market value and revenues during the last few years. Some
of these companies have also become ad hoc monopolies in their
market segments (e.g., Google, Facebook, and Netflix) by acquisitions of
competitors. These companies also benefit from strong network effects,
thereby resulting in robust lock-in barriers for users (see ► Chap. 12).
The deregulation of telecommunications has also generated a new
form of competition in the global telecommunications industry. Until
1998, the old monopolies existed within a single country, but after
1998, these companies could also start operations in other countries.
Making the situation even more complex, two new types of operators
have arrived: resellers and virtual network operators.

5.5 Resellers and Virtual Network Operators

Resellers and virtual network operators are two stakeholders in the
telecommunications market that are direct results of the de-
monopolization of this business area. These concepts are defined as
follows.
 Definition 5.3
The reseller buys bulk traffic from other network operators and
resells it to its own customers.

Definition 5.4
A virtual network operator (VNO) does not own its own access and
network infrastructure but uses the infrastructure owned by other
network operators.

Definition 5.5
The mobile virtual network operator (MVNO) is a VNO offering
mobile services.

The resellers buy bulk traffic capacity and call time from
telecommunications carriers and resell it to their customers with profit.
Reselling is particularly popular in the mobile market. The reseller does
not own any network infrastructure. In the mobile market, they may
issue their own SIM. The profit is generated from discounts they obtain
by buying large quantities of traffic capacity and by combining
telecommunications services with other services or goods, e.g., service
packaging, price profiles, and value-added services. The reseller is the
single point of contact for their customers independently of the
operators from which the reseller buys traffic capacity. The resellers are
in control of their own systems for customer care, billing, marketing,
and sales, either owning these facilities themselves or outsourcing
them to specialized providers of such services.
The mobile market was opened for resellers in Europe in 1992, just
after the first GSM network was put into operation.
The virtual network operator (VNO) buy access to the network
infrastructure of network operators (NOs) owning their own network.
The most common VNOs are the mobile virtual network operator
(MVNO). They deliver their services to their customers using the radio
network infrastructure of mobile network operators (MNOs) owning
base stations and other mobile network infrastructure. The MVNO
issues its own SIMs, operates its own Home Subscription Server (HSS)
for subscription and location management, and has at least one
Internet gateway router and/or telephone gateway exchange for access
to the network of the MNO. The configuration is shown in ◘ Fig. 5.3 for
an MVNO offering 4G services. Data packets from the mobile terminal
are then routed from the base station via the gateway router (GW) into
the Internet, and data packets coming from the internet are routed to
the gateway router before they are routed into the network of the MNO
and delivered to the mobile terminal over the base station.

Fig. 5.3 Network with MVNO. (Authors’ own figure)

What makes the MVNO different from a reseller is that the MVNO
owns some network infrastructure, while the reseller does not. The
actual MNO serving the MVNO is not visible for the customers, and the
MVNO has roaming agreements with other MNOs independently of the
MNO serving the MVNO.
MVNOs are particularly interesting because there are so many of
them. The first MVNO (Sense Communications) was established in
Denmark in 1997 and in Norway and Sweden in 1999. In 2014, there
were 943 MVNOs worldwide (The global MVNO landscape, 2012–2014,
2014).
 The number of MNOs in a region is limited by the amount of radio
spectrum available, and the dominating mobile operators in EEA are
obliged by EU directives to offer services to both resellers and MVNOs
to enhance competitions in the mobile market. The effect of
competition between MNOs and MVNOs may not be obvious as
illustrated in the example below.

Example. Competition and Mobile Virtual Network Operators

Initially, there was strong resistance from mobile network operators
(MNOs) to allow virtual mobile network operators (MVNOs) into their
networks. They were afraid of increased competition without really
appreciating the difference between market share and revenue share.
The size and value of mobile operators are measured in terms of
market shares and not in terms of revenue shares.
◘ Figure 5.4 shows the case of two competing network operators
(MNO1 and MNO2) and an MVNO leasing infrastructure from MNO1.
The MVNO pays a leasing fee to MNO1 for using its infrastructure.

Fig. 5.4 Competing network operators. (Authors’ own figure)

The effect of the MVNO is illustrated by the following simple

numerical example, illustrated in ◘ Fig. 5.5: Suppose that the market
consists of three million subscribers and is equally shared between
the two MNOs before the MVNO enters the market. The revenue per
user is 1000 money units. Then the revenue for each of the two MNOs
will be 1.5 billion money units initially.

Fig. 5.5 Market shares and revenues for mobile network operator (MNO) 1 and 2 and mobile
virtual network operator (MVNO). (Authors’ own figure)

At some time after the MVNO entered the market, the two MNOs
and the MVNO have one million subscribers each generating 1000
money units each; that is, MNO1 and MNO2 have both lost 0.5 million
subscribers to the MVNO. Suppose next that the rent the MVNO has to
pay for using the network of MNO1 is 500 money units per subscriber,
then the revenues of MNO1 will be 1 billon from own subscribers plus
0.50 billion from the MVNO; that is, the revenues of MNO1 is 1.5
billion money units. The revenues of MNO2 are 1 billion money units.
 Compared to the situation that existed before the MVNO entered
the market, the revenues of MNO1 have stayed the same, while the
revenues of MNO2 have become 0.5 billion money unit smaller. This
simple example shows that even if the MVNO is winning many
customers from MNO1, housing the MVNO may still be a good
business for MNO1 since a large proportion of the revenues of the
MVNO is fed back to MNO1 in the form of network leases. Some of
these revenues are revenues lost by MNO2 to the MVNO. The result is
that MNO2 always loses both market share and revenues.

5.6 Conclusions
Telecommunications was a monopoly business up to about 1985: the
operator owned and sold or rented out the user equipment, owned the
network, and offered all services available on the network. Thereafter, it
took about 10 years until the telecommunications market was
completely liberalized. After 1998, the telecommunications market has
been broken up into three major sectors:
– Development and retail of user equipment.
– Network operation.
– Service, applications, and content provision.
There is full competition within each of these sectors. However,
network operations are subject to certain regulation to ensure
interoperability between customers subscribing to different networks.
It is expensive to build and operate telecommunications networks.
Therefore, two types of operators not owning network infrastructure
have been allowed into the market (in particular, the mobile market):
resellers of traffic and virtual network operators. These operators
increase competition and make the market richer for the users.

Questions
1.
What is the key source of revenues for resellers?  
2.
What are the sources of revenues for mobile network operators?  
3. Why may renting out a communication network be a good
 business?
 
Answers
1.
Discounts for buying large amounts of traffic time from ordinary  
operators.
2.
Subscription fees, traffic fees, sales of bulk traffic to resellers,  
rents from virtual network operators.
3.
Increased revenues due to rental fees, and that competitor users  
churn to the MVNO renting the network.

References
Lando, S. D., et al. (1994). Fordham Law Review, 62.

The global MVNO landscape, 2012–2014. (2014, 12 June) GSMA intelligence.

Footnotes
1 PTT = Post, Telephone, and Telegraph

 
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_6

6. Digital Goods and Services

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Digital services – Marginal cost – ARPU – Commodity –

Transaction costs – Service bundling

Learning Objectives
After completing this chapter, you should be able to:
– Explain the concepts of marginal cost, exclusivity,
commodities, and transaction costs for digital services.
– Explain why some digital goods and services can be offered
free of charge.
– Understand why service bundling is particularly simple for
digital goods and services.

6.1 Definitions
Everything in the digital economy can be mapped down to the
production, circulation, trade, and use of digital goods and services. The
definitions of digital good and digital service are related and sometimes
overlapping concepts but are also different in several aspects.
 Definition 6.1
A digital good is a networked zero marginal cost virtual object
having value for some individuals or organizations.

A digital good has, in addition, the following properties (Fournier,

2014):
– The virtual object is intangible but can be stored as data on a digital
medium, e.g., the consumer’s hard disk or smartphone or in the
cloud.
– The virtual object can be replicated without any incurred cost; that
is, the marginal production cost of the object is zero as explained in
► Sect. 6.2.
– The format of the virtual object must be such that it can be delivered
to consumers over the Internet, or in other words, the virtual object
is networked.
– The virtual object must have financial, psychological, or other value
for the consumer (individuals or organizations). Virtual objects
without value for anyone are not included in our definition.
Examples of digital goods that satisfy this definition include
Microsoft Word documents, music tracks on Spotify, webpages on the
Internet, apps on iPhone, Wikipedia articles, e-mails, data stored on
electronic bank accounts, private data stored on Dropbox accounts, and
the list of apartments on an Airbnb web page. These goods are all
virtual objects; they have value for someone; they can be replicated
without any cost; and they can be delivered to consumers over the
Internet. Examples of non-digital goods are computers, mobile phones,
and mobile base stations. These goods have value for someone, but
none of them are virtual objects; they have non-zero marginal cost and
cannot be sent over the Internet.

Definition 6.2
A digital service
is a networked zero marginal cost service that has
value for individuals or organizations.
Services are intangible by nature. Digital services include posting news
on social media, electronic banking, Internet access, multiplayer online
gaming, web browsing, and composing and sending e-mails. The
difference between a digital good and a digital service is somewhat
blurry. This is illustrated by two examples. The data on a Facebook
account is a digital good, while the use of Facebook for any purpose is a
digital service. Music tracks stored on Spotify’s servers are digital
goods, while the use of Spotify to listen to music is a digital service.
Network access and transmission of data over fixed and mobile
networks, as well as data storage and data processing, are also digital
services. These services are not only digital services in their own right
but are also enablers for other digital goods and services. As such, they
are also called enabling or fundamental technologies. This means that
the value proposition of all other digital goods and services depend
critically on these technologies. Moreover, the providers of digital goods
and services usually benefit from the enabling technologies without
investing directly in them. For example, Facebook uses the worldwide
Internet to support its value proposition but has not contributed to the
development and management of the Internet as such. Offering
enabling services as a commercial product has become a new business
arena referred to as Anything as a Service (XaaS), also known as cloud
computing. Cloud computing and XaaS were described in ► Sect. 4.​4 in
the context of business ecosystems. ► Box 6.1 lists some examples of
XaaS.
Digital goods and services are different from physical products (or
tangible goods). Physical products have presence in the physical
domain, while digital goods are built up by sequences of bits and exist
only as pieces of software or data stored on computers or other storage
devices. Digital goods and services may be combined to form larger and
more complex digital goods and services before being offered to
consumers. One simple example is the use of the secure digital payment
solution offered by Google Play where the same platform is used to
support the app software, to pay the developer for use of the app, and
to collect payments from the user. A similar but more complex example
is to use smartphones as authentication tokens for secure access to
bank accounts (see ► Box 6.2). This case involves the primary service
provider (the bank), the authentication provider (the mobile operator),
and one or more clearinghouses which supervise and guarantee the
validity of the authentication process. This example illustrates that a
new digital good or service may be constructed by linking digital goods
and services delivered by several independent stakeholders, in this
case, banking, mobile operation, clearinghouse technology, and Internet
operation. The ecosystem of the new product may then become
complex consisting of elements from the ecosystem of each stakeholder.
Digital goods and services do not degrade as a function of time—
they will continue to exist far into the future with the same quality as
when they were created because of their digital nature. This is different
from physical goods which normally degrade over time.
For simplicity, we will hereafter use the term “digital goods” instead
of “digital goods and services” if the meaning is clear from the context.

Box 6.1 Anything as a Service

The XaaS concept is a generic name for several commercially
available types of enabling services. The most common are Software
as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure
as a Service (IaaS). Different combinations of enabling services are
illustrated in ◘ Fig.  6.1.

Fig. 6.1 Examples of XaaS. From on-premises systems where everything is self-managed
(left) to SaaS where only data is self-managed. (Authors’ own figure)
 Dropbox is an example of an IaaS offering storage of data for
anyone in one of their data centers. The user can access the data via
the Internet from anywhere and at any time.
App Engine of Google is
a PaaS platform where app developers can develop software for the
web applications. Vortex is an SaaS offering gaming services. Using
Vortex, the player need not install, store, or process the game in the
player’s own computer since all processing is done in Vortex servers.
Blockchain technology is also offered as a service called
Blockchain as a Service (BaaS). Consumers buy access to BaaS
without installing the complex processing software needed to
support the service. Amazon, Oracle, and IBM are examples of
companies offering BaaS.
XaaS
has changed the way in which companies invest in ICT.
Companies buying services from an XaaS provider need no longer
bind capital on long-term investments since the use of XaaS converts
these investments into short-term running costs. The capital costs of
offering XaaS may be huge since XaaS usually requires large
investments in computing infrastructure and support of fast and
reliable communication networks to handle many simultaneous
customers and process huge amounts of data.

Box 6.2 E-Banking

E-banking (online banking or Internet banking) is the provision of
digital banking services using the Internet and mobile devices. The
consumer can access their bank data and perform bank operations at
any time and from any geographical location. E-banking has
disrupted the traditional banking services by providing permanent
access to the bank from anywhere and at reduced transaction costs.
Some banks are “virtual banks,” meaning that they only offer online
e-bank services and are only present on the Internet.
For an e-bank, it is essential to provide strong security measures
including the use of strong cryptographic protocols for both message
transfer and user authentication. To provide a complete e-banking
service, cooperation between stakeholders at different business
levels is necessary. This is illustrated in ◘ Fig.  6.2. The user is
accessing their e-bank application through their laptop over the
Internet and is authenticated via their smartphone. In Norway, this
authentication service is called Mobile Bank ID.

Fig. 6.2 E-banking. (Authors’ own figure)

The authentication of the user takes place over the

GSM/3G/4G/5G
network by first authenticating the smartphone of
the user and, thereafter, sending a onetime password via SMS. The
onetime password is then returned to the bank from the laptop to
complete the authentication process. The independent stakeholders
are the bank, the authentication provider, the mobile network
operator, and the Internet provider. Note that cooperation between
different business layers and within the same business layer is
needed to provide a complete e-banking service. For example, mobile
authentication may take place over two mobile networks: the home
 network of the user and a visited network if the user accesses the
bank from another country.

6.2 Zero Marginal Cost

A key characteristic of digital goods is that they have zero marginal
cost. Digital goods can be reproduced with no additional cost.

Definition 6.3
Marginal cost is the cost of producing one additional unit of a good
or service. Digital goods and services have zero marginal cost.

Let us estimate the cost of producing one additional copy of an app. If

the app has been downloaded and installed, say, 10,000 times on
different smartphones, what is the cost of the next download? In a
networked system, the app software is downloaded from the app server
and transmitted over the Internet to the smartphone where it is
installed. There are certain costs associated with these operations, for
example, a small amount of electric energy is required for processing
the app in the server and installing it in the smartphone. Similarly, there
are some small costs associated with sending the app software over the
Internet (amount of electric energy consumption per message and the
share of fixed costs associated with sending a single message over the
network). These costs are tiny so that, for all practical purposes, the
cost of installing one additional copy of the app is zero.
Repeating the same arguments, we again conclude that the costs of
processing an extra trade transaction on Amazon, posting a new
message on Facebook, or downloading a video on YouTube are zero.
On the other hand, the marginal cost is not zero for physical goods.
The cost of production then includes the cost of raw materials (e.g.,
steel or plastic), workforce, and logistics. Note that some e-commerce
businesses selling physical products online, such as Amazon, do not
have zero marginal cost because of the shipping of the physical goods.
However, the marginal cost of the digital trade transaction is zero.
Digital goods may have high fixed costs since some of them are
expensive to develop, build, and operate. The development of software
may cost millions of US dollars, require development teams with
hundreds of people, and take several years from idea to finished
product. The development of the computer game Grand Theft Auto V
(released in 2013) cost $265 million, while Star Wars: The Old Republic
(released in 2011) cost more than $200 million (List of most expensive
video games to develop, 2020). On the other hand, there are examples
of digital goods with low fixed costs, for example, the computer game
Minecraft
and many apps for iPhone or Android. Some of them are very
simple digital services that have been developed by a single person
over a relatively brief timeframe.
An interesting example is telecommunications networks. It costs
billions of US dollars to build and run these networks. On the other
hand, the networks transport several exabytes (1018 bytes) of
information each second so that the cost per byte is negligible, several
magnitudes smaller than a cent. Hence, the marginal cost for sending
an additional message over the expensive telecommunications
infrastructure is, for all practical purposes, zero. The marginal cost of
telecommunications services is thus zero, and the pricing of
telecommunications services cannot be based on usage.
High fixed costs do not always mean that the consumer must pay for
the good. This depends on the business model of the company. The
running costs of Facebook and Google are very high. These costs are not
covered by direct payments from the consumers but, indirectly, by
selling advertisement space. This is made possible by collecting huge
amounts of personal data from the users and selling these data to
marketers for targeting the advertisements or to other data processing
corporations using them for other purposes. Such business practices
are subject to obvious privacy concerns.
The relative gap between fixed and marginal costs is descriptive for
digital goods. For example, the major cost of an app is the development
of the app. The time taken to develop an app can be anything from a few
days to several years. As we have seen, the marginal cost of a digital
good (e.g., the app) is zero so that the cost of installing the app is also
zero, and every new sale of the app contributes directly to revenues.
Copies of the app can even be distributed for free without any financial
loss for the developer. When total sales have matched the costs of
developing the app, the remaining sales are pure profit.
 Under these conditions, the average cost of a copy of the digital
good equals:

where AC is the average cost, F is the fixed costs, n is the number of

copies of the good produced during its lifetime, and MC is the marginal
cost. ◘ Figure 6.3 shows the average cost as a function of n. Observe
that the average cost approaches zero as n gets large.

Fig. 6.3 The average cost as a function of the number of units produced. (Authors’ own figure)

Economies of scale are the cost advantages that companies obtain as

the number of units produced increases. This is so because the fixed
cost per unit decreases as the production volume increases as shown in
◘ Fig. 6.3 and may tend to zero if the number of units produced is large.
Companies producing physical goods benefit from cost advantages as
the number of units produced increases only up to a certain point. The
marginal cost per unit produced is independent of production volume
and sets a lower limit for the cost of the product. Moreover, expanding
the production beyond a certain threshold may also necessitate that
more production infrastructure must be built, thereby increasing the
cost of administration and support-functions so that the benefits from
economies of scale are marginalized. This is different in the digital
economy. This is so because the marginal cost is zero and that there is
no limit to the number of units that can be produced without increasing
the fixed costs. Hence, the cost per unit produced will be zero
independently of the production volume. This is one of the reasons why
companies producing digital goods and service get so big.

6.3 Classification of Digital Goods

Goods can, in general, be classified either as private goods, public
goods, club goods, or common-pool resources (Stiglitz, 2015)
. This
classification depends on whether a certain good is rival/non-rival or
excludable/non-excludable.

Definition 6.3
A good is classified as rival if it is reduced in quantity after
consumption or if the usage of the good prevents others from using
it. A non-rival good is the opposite of a rival good; it is neither
reduced by consumption nor does the usage of the good prevent
others from using it. An excludable good is such that it is possible to
prevent consumers from accessing or using the good. A non-
excludable good is such that consumers cannot be prevented from
accessing or using the good.

From these characteristics, four different types of goods can be defined

as shown in ◘ Table 6.1.
Table 6.1 Different types of goods. (Authors’ own compilation)
  Rival Excludable Examples
Private goods Yes Yes Cars, food, mobile phones
Common-pool resources Yes No Fish, forests, wild berries
Club goods No Yes Cinemas, cable TV, private parks
  Rival Excludable Examples
Public goods No No Air, national defense, knowledge

Digital goods are non-rival by nature—the consumption of a digital

good by a user does not reduce the quantity available to other users of
the same digital good. For example, a user reading a webpage does not
reduce the availability of that webpage for other users. A Spotify
subscription gives a user access to Spotify, but this does not prevent
other users from accessing Spotify. A user accessing the Internet does
not reduce the availability of the Internet for other users. The latter is
true with some limitations since webservers and the Internet have a
maximum capacity. However, most computer and communication
systems today are provisioned to handle high demands. In this book,
digital goods and services are assumed to be non-rival in most practical
cases.
Digital goods can either be excludable or non-excludable.
Excludability means that access to the good can be regulated. On the
other hand, if a good is non-excludable, a user cannot be denied access
to the good. Digital goods that are widespread on the Internet are non-
excludable. Examples are free music, news, and content on free web
pages. Digital goods that have restricted access are excludable.
Examples include access to specific magazines and journals,
copyrighted music and movies, and licensed software. Excludable goods
can be accessed by, for example, accepting a paid subscription plan or
enjoying a club membership. The illegal copying of copyrighted
material might result in excludable content becoming non-excludable—
copies become abundant and available for everyone.
Digital goods and services are non-rival by nature. Digital goods
and services can be either excludable or non-excludable.

◘ Figure 6.4 shows examples of digital goods classified according to

the type of good defined in ◘ Table 6.1. Note that all the digital goods in
the example are non-rival. Access to Wikipedia articles is non-
excludable since the website is open and available for anyone. Gmail is
an open and free service available for anyone who registers for an
account. Spotify is both excludable and non-excludable at the same
time. Its basic service is free for anyone registering for Spotify;
however, Spotify’s premium service is accessible only by paid
subscription. Internet access is excludable since the users must pay for
Internet access. However, Internet access may also be non-excludable;
access to the Internet may be free and open to anyone in airports and
shopping malls and on trains and other public transport. Access to
Netflix and World of Warcraft
is excludable—a paid subscription plan is
required for using both these services.

Fig. 6.4 Classification of digital goods. (Authors’ own figure)

If a digital good is both non-rival and non-excludable, it is classified

as a public good. Public goods occupy a special place in economic
theory. They are often prone to market failure resulting in
governmental regulation or other market interactions to make the
market function properly. One of the reasons for market failure is the
free rider problem. Public goods can be accessed by anyone without
paying for it. There are, therefore, no incentives for private actors to
provide these goods; likewise, there are no incentives for users to pay
for access or usage to maintain and uphold the good. One example is
Wikipedia offering a free encyclopedia on the Internet. There are few or
no incentives for private actors to financially support or invest in
Wikipedia. Wikipedia relies primarily on two sources of income for
sustaining its operations: financial donations from benefactors and
voluntary work by authors who write and update articles.
A common-pool resource is prone to the tragedy of the commons
, in
which common resources get overused and depleted. Even though
individuals act in their own self-interest, it is the collective behavior of
all individuals that may lead to overuse or depletion. Examples of the
tragedy of the commons are overfishing in the seas and air pollution.
The tragedy of the commons also has impacts on the digital economy
even though digital goods and services are non-rival by nature. Since
digital goods and services have zero marginal cost and, therefore, close
to unlimited supply, the tragedy of the commons takes different
appearances in the digital economy, resulting in publishing of unwanted
and illegal information and content, spam, denial of service attacks, and
service abuse (Gapper, 2017).

6.4 Zero Average Revenue per User

Average revenue per user (ARPU) is an important financial indicator in
economics. The ARPU states how much revenue an average user is
generating over a specific period of time. For instance, the monthly
ARPU for mobile subscribers in the USA is about $40. Total revenue for
a mobile service provider is the ARPU multiplied by the number of
subscribers. The mobile service provider will focus on increasing its
total revenue by increasing both the ARPU and the number of
subscribers.
In the digital economy, several companies operate with zero
average return per user (ARPU = 0), in which case the company does
not receive any revenue from the consumers at all. Since the marginal
cost is zero, the cost to attach users is also zero. This applies to all
competitors offering the same service so that, to compete, goods are
offered for free to the consumer by all of them. The suppliers do not
compete on price but on user experience. In this economy, supply-
demand curves are meaningless.
The main challenge for many companies in the digital economy is,
therefore, to get revenue for its operations from sources other than
consumers. A few companies have succeeded in this effort, for example,
Facebook, Google, and Twitter. Their main source of revenue is
advertisements: the more they know about their consumers, the more
attractive they are as marketing channels.
Note that ARPU = 0 is not a universal rule applicable to all digital
services. Netflix, for example, has an ARPU >0 because consumers pay
to access the service.

Box 6.3 Facebook and ARPU = 0

Facebook is the world-leading social networking service with
currently more than two billion users. None of the users pay anything
for using their services so that ARPU = 0. Yet, Facebook’s revenue is
over $40 billion for 2017, increased from $27 billion in 2016 and $18
billion in 2015. In 2018, Facebook is among the six most valued
companies in the world, according to market capitalization (◘ Fig. 
6.5).
 Fig. 6.5 Facebook revenue and market cap 2012–2020. (Authors’ own figure)

How is it possible for Facebook to reach such a financial position

when it does not earn anything from its users? The answers to this
question are related to network effects and multi-sided platforms
(see ► Chaps. 9 and 10).

6.5 Digital Commodities

Commoditization is the process by which digital goods (or any good or
service in general) end up being indistinguishable from a consumer’s
point of view. Competing goods will look the same to the user—it is
impossible to differentiate between the goods even though they are
produced by different manufacturers. The only distinguishing factor for
a commoditized good is the price. It is, for example, impossible for
consumers to distinguish between lubrication oils from different
refineries or electricity produced by different power plants—only the
price can be used as a distinguishing factor.
 Several, but not all, digital goods have been commoditized.
Examples of digital commodities are Internet access and transport of
bits, storage of data, processing of data, international news bulletins,
and, to some extent, certain types of software products (e.g., word
editing and spreadsheet software).
Digital goods that have been commoditized compete only on price.
A fierce competition among companies providing digital commodities
tends to push the price to zero because of the zero-marginal cost
property of digital goods. Standard microeconomic theory on perfect
markets also predicts this outcome. However, at price equal to zero, it is
a challenge for companies to be profitable. Most of them will run out of
business as revenues decrease and profits turn negative. This is a
strategic dilemma for several companies in the digital economy. To
avoid a price war resulting in zero revenue, companies may
differentiate their goods from the competitors by making them unique
for customers, for example, newspapers offering customized news
alerts. The newspaper may then generate revenue from these
customers. Another option is to apply creative business models that
operate under condition that the service is provided for free while the
revenues are obtained from other sources—Google and Facebook
employ such business models.
Some digital goods become commodities because of standards.
Standards force providers to deliver indistinguishable digital goods, for
example, compatible Bluetooth access to mobile phones, hearing aids,
and household gadgets. Digital goods also tend to become commodities
over time. This may be the result of the harmonization of competing
goods. An example is word editing software. In the 1990s, editing
products from different suppliers supported different functions. Today,
word editing software packages are so similar that they have become
digital commodities.

6.6 Transaction Costs

The transaction cost is the cost associated with the process of selling
and buying. Transaction costs can be divided into three categories:
– Search and information costs are the costs of searching for a
particular good and determining its price and properties.
– Bargaining costs are the collective costs for the consumer and the
provider to agree on the terms of the contract. This includes the
price and delivery conditions of the good.
– Policing and enforcement costs are the costs of sticking to the
agreement and taking appropriate action if the agreement is not
upheld by either party.
An example is trading of physical goods on eBay. The buyer first
searches for products they want to buy among offers made by various
sellers (search and information costs). This is done on eBay’s website.
When the buyer has decided which item to buy, buyer and seller will
negotiate price and delivery conditions (bargaining costs). If something
is wrong with the received product, the buyer will enforce contractual
rights by taking direct contact with the seller (policing and
enforcement costs). eBay may also be involved in the dispute if the
buyer and seller do not agree.
Amazon
, eBay, and Alibaba lower the transaction costs by offering
efficient search processes using web or mobile apps.

6.7 Bundling
Product or service bundling means that several products or services are
combined and offered for sale as a single package. One example is the
Microsoft Office 365 package. This package contains the digital services
Word, Excel, PowerPoint, OneNote, Outlook, Publisher, and Access.
Another example is cable television subscriptions where the user may
subscribe to various bundles of television channels. The cable television
provider may also extend the package to include audio broadcasting,
streaming of movies and music, broadband Internet access over Wi-Fi,
and VoIP. This is exemplified in ◘ Fig. 6.6.
Fig. 6.6 Example of a bundle of the services Netflix, Spotify, and network access. (Authors’ own
figure)

Pure bundling means that consumers can only buy the bundled
package without the opportunity to buy the single products or services
in the package. Mixed bundling means that a consumer has the option
to buy the package as well as the single products or services
constituting the package. In general, the price of the bundle is lower
than the sum of the price of the individual services constituting the
package.
Bundling is a strategy for providers to increase sales. In the digital
economy, bundling is common and particularly efficient because of the
zero marginal cost property—it does not cost the provider anything
adding another service it already owns to the package. The consumer
may find this business model attractive since an additional service in
the bundle contributes to increased value for the consumer and, hence,
an increased willingness to pay for the bundle.
The digital economy has also enabled the unbundling of previously
bundled goods and services. Unbundling has become common in the
music industry. Apple iTunes, for example, unbundles CD albums
offering each music track as an independent product. The customer
may then buy a single track rather than buying a complete album. It
also allows the consumer to build their own personal playing lists
containing songs from several albums. The benefit for the music
industry is that this increases sales without generating additional costs
since the marginal cost of providing a complete album or just a single
track is zero in both cases.

6.8 Conclusions
The most important characteristics of digital goods and services are
summarized as follows:
– The marginal cost of production and distribution of digital goods and
services is zero.
– Digital goods and services are non-rival; that is, the availability is not
reduced by consumption and usage.
– Digital goods and services may be excludable or non-excludable
depending on commercial conditions of usage.
– Several, but not all, digital goods and services are commodities; that
is, they can only be distinguished by price. In some market segments,
the price for the product is zero, and price cannot be used as
differentiating factor.
– It is easy to bundle digital goods and services in flexible packets. If
the provider owns all components of the bundle and all components
are digital goods, then the cost of bundling is zero. In this case, the
customers may create their own bundles, for example, playing lists
for music and collection of television channels.

Exercises
1. Digital Services
Which of the following are digital goods or services?  
1.
The New York Times
web page  
2.
Internet access over a 4G mobile network 
3.
A compact disc (CD)  
4.
HBO subscription  
5. An Apple iPhone
  
2.
Computer Game Development  
A computer game costs $100 million to develop (fixed costs).
The game is sold to consumers of $50 per copy. Assume that the
marginal cost is zero. How many copies must be sold to cover
the fixed costs? How many copies must be sold to cover the fixed
costs if the marginal cost of each copy is $10? Under what
condition is the marginal cost zero? Under what condition is MC
> 0?
3.
Digital Commodities  
Determine if any of the following digital goods have been
commoditized:
1.
Social media services  
2.
E-mail clients  
3.
Web browsers  
4.
Cloud storage  
5.
Wi-Fi
Internet access  
Why have these digital goods been commoditized? If
they have not been commoditized, why not?
4.
Bitcoin
Transaction Costs  
How does Bitcoin
influence transaction costs in the digital
economy? Which stakeholders are affected by Bitcoin’s potential
impact on transaction costs?
5. Internet Explorer
In 2002, Internet Explorer had over 90% of the web browser  
market. Who were Internet Explorer’s main competitors at that
time? Explain how Internet Explorer got so large.
Answers
1.
The following are digital services: The NY Times web page,  
Internet access, and an HBO subscription. The following are not
digital services: a CD and an iPhone. The NY Times web page is
considered a digital good. The content available on HBO is a
digital good.
2.
The number of copies sold must cover the fixed costs. If MC = 0,  
the number of copies sold to cover the fixed costs is 100,000,000
/ 50 = 2,000,000 copies. If MC = 10, the number of copies sold to
cover the fixed costs is 100,000,000 / (50 - 10) = 2,500,000
copies. The MC = 0 if the trade is completely digital. That is,
ordering, payment, and delivery are all digital. On the other
hand, the MC > 0 if the trade includes a physical product (e.g.,
DVD containing the game, a box, or a user manual).
3. Of these services, only cloud storage is fully commoditized.
Social media is not commoditized. This is because the social  
media networks available (e.g., Twitter, Facebook, and Google+
)
have very different functionalities, and they do not communicate
with one another. This is because of missing standards in the
social media industry. E-mail clients are close to being
commoditized. This is because e-mail as a service is
standardized (SMTP
). Hence, the functionalities of different e-
mail clients are the same. The major difference between various
e-mail clients is the design and interoperability with other
applications (such as a calendar). Web browsers are close to
being fully commoditized. This is because the web service is
standardized (through the protocols HTML, HTTP, and URL).
The major differences between web browsers are the design
and, to some degree, the interoperability with other digital
services (such as Flash and JavaScript). Cloud storage is fully
commoditized. The only difference between various cloud
storage services is their price. Wi-Fi Internet access is close to
being fully commoditized. However, there are quality differences
in Wi Fi access (e g different bandwidths)
 in Wi-Fi access (e.g., different bandwidths).
4.
 
Bitcoin
has the potential to significantly reduce transaction costs
in the digital economy. Current third-party payment services—
such as VISA, MasterCard, and American Express—charge
transaction fees of 1.4–3.5%. Bitcoin, on the other hand, can
charge as little as 0%. It is no wonder why the established
financial industry views Bitcoin and other cryptocurrencies as a
threat to their operations. It is VISA, MasterCard, American
Express, and, to a certain degree, the banking industry, which
are the competitors of Bitcoin.
5.
The main competitor of Internet Explorer was Netscape.  
Netscape was the largest web browser in the mid-1990s.
However, when Internet Explorer was launched in 1995, it
started to increase its market share until it had almost a
monopoly in 2002. The main reason for Internet Explorer’s
success was because it was free and bundled with Microsoft
Windows. Few consumers took the extra trouble of installing
another web browser on their PC when Internet Explorer was
already installed and ready for use. This, combined with the
large market share of PCs using Microsoft Windows, fueled
Internet Explorer’s growth. In addition, Microsoft—which
provided Internet Explorer—was a much larger company than
Netscape, with significantly more financial resources.

References
Fournier, L. (2014). Merchant sharing: Towards a zero marginal cost economy. ArXiv, 1405.2051.

Gapper, J. (2017, November 29). Facebook faces the tragedy of the commons. Financial Times.

List of most expensive video games to develop. (2020, December 16). Wikipedia.

Stiglitz, J. E. (2015). Economics of the public sector. W. W. Norton & Company.

Further Reading
Christensen, C. (1997). The innovators dilemma: When new technologies cause great firms to fail.
Harvard Business School Press.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_7

7. Production Models
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords In-house production – CBPP – Crowdsourcing –

Crowdfunding

Learning Objectives
After completing this chapter, you should be able to:
– Comprehend the difference between production of physical
and digital goods.
– Explain the basic production models described in this chapter
and apply them in the analysis of business models.
– Understand the importance of online financing methods and
open-source software for the development of digital goods.

7.1 Physical and Digital Products

Production of digital goods and services is different from standard
industrial production. Standard theory of production is based on
manufacturing of physical goods requiring, in addition to the actual
manufacturing process, retrieval of raw material, storage facilities for
products before they are delivered to the market, and logistics for
transporting the products to the market. All the elements of the
production process contribute to the total cost of the product as
explained in ► Sect. 8.​2. The costs associated with development and
later improvements of the product are usually not included as a
contribution to the direct costs of production. These costs, together
with administrative and capital costs, are usually included among the
common costs of running the company. The total direct cost of
production is then equal to the marginal cost of producing one item of
the product multiplied by the number of items produced.
Production of digital goods is different, though all the production
stages mentioned above may be present. The most important
characteristics of the production of digital goods include:
– The major cost contribution to a digital product is associated with
developing the product itself (e.g., app, webpage, book, or program
update) and investments and running costs associated with the
production platform, for example, fees to cloud providers. Note that
the running cost for producing some digital goods may be huge.
Production of Bitcoin requires huge computer facilities consuming
large amounts of energy. The same may apply to search engines and
providers of massively multiplayer online games running large
server platforms capable to accommodate millions of simultaneous
players.
– After the product and the production platform have been developed,
the cost of producing an item of the product is almost zero (zero
marginal cost).
– Only one copy of the product needs to be stored. Copies of the
product are produced instantly on demand.
– Distributing a single copy of the product over the Internet costs
almost nothing.
– Some digital products may require “raw materials” in the form of
licenses or other expenditures associated with each item, for
example, copyrighted material such as music, books, and films. For
these products, the marginal cost is no longer zero.
The difference between properties of digital and physical products
is explained in ► Chap. 6.
The general characteristics of digital goods generate new ways of
production. The most important difference compared to physical
products is that digital products can be produced by collaboration over
the Internet giving rise to entirely new production methods. In the
following, production is understood to include both the development of
the product and the subsequent production and delivery of the product
to the customers.
There are several ways in which digital goods can be developed
and produced. Three basic production methods of digital goods are
in-house production, commons-based peer production (CBPP), and
crowdsourcing. Production based on CBPP and crowdsourcing takes
place over the Internet.

The three basic production models are not independent—a

company may apply a mix of them when developing and producing a
digital service. Within this context, note that some digital services may
be rather simple to develop requiring only few resources. This includes
even big services like the World Wide Web, Facebook, Airbnb, and
TCP/IP
. However, other digital services require the collaboration of
hundreds of people over long periods of time. Examples include the
development of operating systems (e.g., UNIX and Linux), public mobile
networks (GSM, 3G, 4G, and 5G), and local access networks (e.g., Wi-Fi,
WiMAX, Ethernet, and Bluetooth)
.

7.2 Basic Production Methods

7.2.1 In-House Production
In-house production (also termed “firm-production”) means that the
company develops the good or service using internal resources,
possibly combined with manpower contracted from other enterprises
(outsourcing). In this case, the company is in control of the whole
development process and owns the final digital product and intellectual
property rights associated with it. This also includes the part
outsourced to independent companies. In-house production is
illustrated in ◘ Fig. 7.1.
Fig. 7.1 The in-house production model. (Authors’ own figure)

In-house production is the most common approach taken by both

large companies (e.g., Google, Microsoft, and Facebook) and small
startups. The company organizes the work and sets up tasks in such a
way that the final service or good is produced within deadline and
according to accepted industry standards. In-house production may
include the use of open-source software and crowdsourcing for part of
the production.

7.2.2 Commons-Based Peer Production

Commons-based peer production (CBPP
)—also called social
production—is a way of producing goods and services in which a large
number of people (collaborators) take part in the development of the
product. The term was coined by Yochai Benkler (Benkler, 2002). The
production process takes place by use of the Internet—the commons in
the context of digital products. The group of collaborators is usually
self-organized (if organized at all) and without central leadership or
coordination. A platform for gathering the contributions from each
collaborator into a final digital service must be established before the
collaboration starts. The platform is used throughout the production of
the digital service to organize and divide work between the
collaborators. The contributing collaborators are not organized by a
firm as in the in-house production model. Often, the collaborators do
not receive any financial rewards for their contribution. ◘ Figure 7.2
illustrates the CBPP model.

Fig. 7.2 The commons-based peer production model. (Authors’ own figure)

The most famous example of the CBPP model is the operating

system Linux in which hundreds of computer scientists contributed to
the evolution of the Linux software over many years. Wikipedia is also
the result of the CBPP model—no one is coordinating the content or the
evolution of the encyclopedia. Arbitrary readers are checking the
validity of the articles, correcting errors, adding novel material, and
writing new articles.
The Internet itself is also a result of the CBPP model. Anyone can
submit proposals for new protocols, procedures, and functionalities of
the Internet in memoranda called Requests for Comments (RFCs). The
work is loosely organized by the Internet Society (ISOC) and the
Internet Engineering Task Force (IETF
), which are both non-profit
organizations in which anyone—individuals as well as industries and
organizations—can be members. The acceptance and implementation
of new proposals depend entirely upon how the business opportunities
and other prospects of the new proposal are assessed by users and
providers of Internet hardware, software, and services. Some of the
new products may be short-lived (e.g., the real-time transport protocol
XTP developed to speed up data exchange in distributed processing
systems), while others may exist for generations (e.g., HTTP and TCP).
Below follow examples of peer production initiatives in four
different fields: software development, production of encyclopedia,
proofreading of scanned books, and open hardware design.

Example 7.1 Linux and GNU

Linux is a Unix-like operating system suitable for a wide range of
computing devices: mainframe computers, smartphones, servers,
personal computers, and several other types of electronic devices. It is
built on free and open-source software. Linus Torvalds started the
development of the platform in 1991. The idea was to develop an
operating system like Unix consisting entirely of free software that
could be loaded down by anyone free of charge. Since then, several
hundred computer scientists have contributed to the development of
the software, and in several countries and regions, there exist loosely
knit Linux User Groups promoting the software, assisting new users
installing the software, and providing technical assistance and
training. Anyone may contribute to the evolution of the software but
without levying any wages for the work.
The goal of the GNU project is to provide an environment for
collaborative software development and distribution of free software.
The distribution of the software is subject to GNU General Public
License (GGPL), allowing users to run, share, and modify the software
(also called copyleft conditions, (see ► https://​www.​gnu.​org/​
copyleft/​). The license guarantees that there are no hidden
intellectual property rights or other restrictions of usage associated
with the software or other products licensed by GNU such as
specifications, design descriptions, and hardware. The Linux kernel
and several other free software packages not developed in the GNU
project itself are distributed with the GNU General Public License.
GNU Free Documentation License (GFDL) is a similar license for free
documentation (e.g., textbooks and manuals) allowing users the right
to copy, modify, and distribute the text. New software or texts
produced and submitted under GGPL or GFDL should also be freely
available subject to copyleft conditions.
The production of the content of Wikipedia is a result of peer
production over the Internet.

Example 7.2 Wikipedia

Anyone may contribute to the content of the encyclopedia by
producing new articles, adding new material to existing articles, and
deleting, modifying, or correcting existing articles. The Wikipedia
Foundation is not a licensing organization itself but a platform on
which contributors can publish their work. The content of Wikipedia
is licensed using the Creative Commons (CC) license. This license is
similar to GFDL but contains rules by which otherwise copyrighted
material (e.g., pictures) can be used in other contexts. The rules may
be different in the various language editions.
By the end of 2020, the English Wikipedia contained about 6.2
million articles and is among the 15 most visited website on the
Internet (See ► https://​stats.​wikimedia.​org/​EN/​TablesWikipediaE​N.​
htm).

Example 7.3 Project Gutenberg and Distributed Proofreaders

Project Gutenberg is digitizing public domain books and other texts
for free public access as e-books. Public domain books are books
where the intellectual property rights have expired or otherwise not
applicable.
As of 2019, the project has digitized about 60,000 books. The
scanning and preparation of the books were first done by the founder
of the project, Michal Hart, alone, but the project attracted soon
several volunteers scanning and preparing books and other texts for
the project. Since Hart died in 2011, the project is run entirely by
volunteers.
When scanning the books, it is likely that errors occur because the
scanner may misinterpret a letter or a group of letters. Proofreading
is required to correct these errors. In 2000, Charles Franks opened
the website Distributed Proofreaders where volunteers could
 proofread the texts produced by the Gutenberg Project. Several
thousand volunteers have since joined the project. Distributed
Proofreaders have prepared about 40,000 texts for Project Gutenberg.

Example 7.4 RepRap Project The RepRap project was initiated by

the University of Bath, England, for developing low-cost 3D printers
using various thermoplastic materials. The project now includes
several hundred participants all over the world developing free 3D
printing software and hardware design and experimenting with new
materials suitable for printing. RepRap is an open design project
releasing design details via the GNU General Public License. This
includes free open-source software, design details, and hardware. The
goal is to develop desktop manufacturing systems affordable and
useful for everyone for, among others, producing household artifacts.

7.2.3 Crowdsourcing
In the crowdsourcing production model, organizations and individuals
produce digital services by inspiring the public to contribute to the
project. This seems somewhat similar to outsourcing; however, the
difference between outsourcing and crowdsourcing is that outsourcing
implies that the work is contracted out to another company on strict
commercial and juridical conditions, whereas crowdsourcing implies
that the work is done by arbitrary groups of individuals without
formalized participation. The result of the work may then be more
arbitrary but may also lead to better, cheaper, and more versatile
solutions. Since there may be many participants in the project, it is
likely that design flaws are discovered at an early stage and that better
and cheaper designs are found. The crowdsourcing production model is
illustrated in ◘ Fig. 7.3.
Fig. 7.3 The crowdsourcing production model. (Authors’ own figure)

Crowdsourcing is a form of peer production. The difference is that

CBPP may follow arbitrary development paths sometimes resulting in a
viable product, whereas a crowdsourcing project aims at developing a
predefined product. Crowdsourcing may be used in any stage of
product development from the initial idea to development, production,
testing, and marketing. The prerequisite is that the collaboration can
take place over the Internet. A company may use crowdsourcing at all
stages of production or only parts of production, leading to a final
digital service.
Next follows some examples of crowdsourcing initiatives.

Example 7.5 Crowdsourcing Using Contests Contests are

important methods to trigger crowdsourcing events. One example is
the CAESAR competition in which the goal was to develop new
encryption methods. The contest started in 2012, and individual
researchers or groups of researchers proposed new algorithms which
were tested against cyberattack and other weaknesses by the
participants in the contest. Six winning algorithms were announced in
March 2019 after several rounds of eliminating proposed candidates.
 The winning algorithms are candidates for new cryptographic
standards.
In 2009, Netflix used a similar contest to improve its
recommendation algorithm. The winner received one million US
dollars (Jackson, 2017). The large premium was then the obvious
motive to participate. Microsoft used crowdsourcing to test the
security of the Windows 8 software. This included both detecting
security bugs ($100,000) and methods to correct them ($50,000)
(Bell, 2013).

Example 7.6 Crowdsourcing Using Collaboration Platforms The

business models of Topcoder and Clickworkers are based entirely on
crowdsourcing.
The core business of Topcoder is to run a collaborative computer
platform and to organize communities of individual coworkers.
Computer scientists, system designers, and software engineers may
join these communities and participate in various projects the
company is developing for their clients. The company also uses online
annual tournaments called Topcoder Open where members of the
community compete in solving particular problems in various fields
of computer science.
Clickworker offers “micro-jobs” to volunteers. Potential
collaborators register in the community run by the company and are
allocated jobs in accordance with their reported skills and earlier
accomplishments. The company pays for the jobs each individual does
for them. These jobs range over a large field of problem areas, for
example, software development, proofreading of books and manuals,
information gathering, and processing of unstructured data. The
participants are connected to the company on a freelance basis, and
they decide themselves when and how much they want to work.

7.3 Production Tools

The most important challenges related to design and production of
digital goods are financing the development and reducing the running
production costs. For new startups, it may be difficult to persuade
banks, venture capitalists, and other investors to put money into the
project. Financing the project may then be achieved by crowdfunding or
peer-to-peer lending. The development costs may be reduced by using
free and open-source software, and investment and running cost may
be reduced by renting processing and storage capacity from cloud
providers.

7.3.1 Crowdfunding
Crowdfunding is a way to raise funds for projects or startups by inviting
people to invest in the venture. The interactions take place over the
Internet. The investors usually receive shares in the startup or company
owning the project equal to the amount of money they have invested.
Most fund-raising events are small but some of them are huge. In 2018,
a Cayman Island startup raised more than 4 billion US dollars for
developing a blockchain platform (Rooney, 2018).
Crowdfunding may refer to several types of funding mechanisms
(Hellmann et al., 2019).
– The most widely used method is peer-to-peer lending (see next).
– Equity-based crowdfunding implies ownership of securities in the
company or project. These are then usually high-risk investments.
– In reward-based crowdfunding, the investors receive services or other
benefits as reward for their investments. In this case, there is no
guarantee that the compensation will match the size of the
investment.
– Donation crowdfunding is mainly used for supporting charitable
causes, and the donor does usually not receive any compensation for
the investment.
There were more than 2000 crowdfunding platforms in 2016
(Drake, 2017). Kickstarter is one of the most successful of these
platforms.

Example 7.7 Kickstarter Kickstarter is an example of a

crowdfunding organization. Project owners publish and describe
their project on the Kickstarter website. People from all over the
world may choose to fund the projects they like, often with as little as
$1. There may be thousands of different investors on a single project.
The project will start if the project achieves its funding goals.
 Kickstarter projects can be anything from game developments and
app design to the creation of books, music, art, and films. All projects
awaiting funding are listed on their homepage. On Friday, January 10,
2020, there were 467,867 projects asking for funding on the
Kickstarter website, 3630 of them were software projects. When a
project is funded and the project owner has completed the project,
the funders will receive a reward based on the allocated funding, e.g.,
a copy of the creative work. ◘ Figure 7.4 illustrates the creative and
funding process of Kickstarter.

Fig. 7.4 The creative and funding process of Kickstarter. (Authors’ own figure)

Kickstarter apply a 5% fee on the collected funding as a source of

revenue. Crowdfunding has been particularly popular in the video
game and traditional table-top game industry, and several games have
been funded using the Kickstarter website.

7.3.2 Peer-to-Peer Lending

Peer-to-peer lending (P2P lending or crowdlending) means that
individuals lend money directly to other individuals or businesses.
Startups may have problems financing the project by traditional bank
loans because of the high risk associated with the project. In such cases,
peer-to-peer lending is an alternative way of financing the project.
Companies mediating in peer-to-peer lending transactions run
match-making platforms where borrowers and lenders may register for
a certain fee. The intermediating company does not offer loans using its
own money and is thus not a regular financial institution.1 The
company may, however, offer services such as credit checking,
registering the transaction details, and assisting in resolving disputes.
The intermediators are value networks as defined in ► Sect. 8.​3,
matching people or firms lacking capital to initiate a project and
potential lenders.
The interest rates on peer-to-peer loans are generally lower than for
bank loans but higher than the interest rates on bank deposits.
Moreover, it has been reported that this market is less volatile than the
stock market, and an investor targeting peer-to-peer lending will
typically diversify the funds over many borrowers to reduce risks and
increase the return on investments (Roth, 2012). People are then
encouraged to invest money in peer-to-peer lending rather than in
stock market securities. The Australian peer-to-peer lending company
SocietyOne reports that they passed $800 million in lending by the end
of 2019.
In some cases, the interest rates are very low and sometimes even
zero, in particular, if the cause is charity or investments in projects in
developing countries (Brook, 2007), for example, using the peer-to-
peer platform of Kiva.
Peer-to-peer lending stands for about 73% of all crowdfunding for
small- and medium-size businesses according to a report from the
Emerge Partnership (ITU) (A review of Micro, Small and Medium
Enterprises in the ICT Sector. Emerge Partnership (ITU). 2016).
Unlike bank loans, peer-to-peer lending is usually not protected by
government guarantees.

7.3.3 Free and Open-Source Software and Cloud

Computing
Free and open-source software (FOSS) is an initiative for the
collaborative development of software. Free and open-source software
is an important ingredient in commons-based peer production. The
source code of the free and open-source software is made available to
other developers along with a license allowing them to use, modify, and
redistribute the software. The license is issued by the copyright holder
of the software and approved by, for example, the nonprofit
organization Free Software Foundation. The GNU General Public
License mentioned in ► Example 7.1 is one such license.
The use of the software may be free or subject to a small fee. There
are thousands of licensed open-source software packages covering
fields such as grid computing, web browsing, file systems, 3D
animation, and firewalls. Linux and Android are among the most
successful open-source projects. In fact, most of the Internet runs on
open-source software.
Cloud computing is also a major component in the production
model of several small- and medium-size enterprises and startups. The
main advantage for a provider of digital services to use cloud
computing is that both investments in computer infrastructure and
running costs associated with operation and maintenance are reduced.
There is also no need for the service provider to have deep insight in
computer science in order to design, promote, and deliver the service.
Other advantages of cloud computing are that it is easier for the
user to adjust resources to fluctuating needs, and the technology offers
increased reliability because data may be mirrored on several
redundant sites, more and better cybersecurity protection, access to
high-speed computing resources with low latency, and state-of-the-art
data speed, streaming capabilities, and protocols (What is cloud
computing? A beginner’s guide. Microsoft Azure).

7.4 Conclusions
The traditional production method is in-house production in which the
entire production of the good takes place within the company. This
method is used for almost all physical products. This is also the
dominating production method in the digital economy. Microsoft,
Facebook, Google, Netflix, and all the other big companies in the digital
economy base their production method on in-house production.
The Internet has created more flexible production methods based
on collaboration over the network. Commons-based peer production
(CBPP) is founded on voluntary participation and takes place without
(or with only little) central coordination. This has led to large projects
such as Linux, digitization of books and documents, development of
free and open-source software (FOSS), and production of technical
standards end recipes. CBPP has created vast amounts of freely
available resources such as statistics, encyclopedia, software, and
digital books and documents.
Crowdsourcing implies that a firm invites individuals to participate
in the development or production of a good. The project is managed by
the firm, but the work is done by a group of arbitrary individuals.
Crowdsourcing does not depend on the Internet but is facilitated by it.
Investments in and financing of new products also take place over
the Internet in terms of crowdfunding and peer-to-peer lending by
bypassing traditional banking and investment procedures.

Questions
1.
Who owns a service produced using the CBPP model?  
2.
How can services produced using the CBBP model generate  
revenues?
3.
What are the main challenges of the crowdsourcing production  
model?
4.
Discuss whether the crowdsourcing production model can  
produce digital services of the same quality as the in-house
production model.

Answers
1.
Anyone who participates in the development of the product may  
be an owner of it. In some cases, the product is marketed and
maintained by a group of volunteers (e.g., Linux, Gutenberg
Project). In many cases, no one owns the product.
2. In most cases, CBPP does not generate revenues since the
product is offered for free. In some cases, the project may
 product is offered for free. In some cases, the project may
receive donations (e.g., Wikipedia) or generate revenues from  
ads, complementary services (e.g., consultancy services or
seminars), or services built on the CBPP service.
3.
Challenges are:  
– Coordinating the work
– Scheduling and planning tasks
– Attracting the right workforce and expertise
– Quality management
4.
Crowdsourcing may produce better results, more innovative  
solutions, and better quality.

References
Bell, L. (2013, June 20). Microsoft offers a $100,000 bug bounty for cracking Windows 8.1. The
Enquirer.

Benkler, Y. (2002). Coase’s Penguin or Linux and the Nature of the Firm. The Yale Law Journal,
112(3).

Brook, Y.. (2007, August 20). The morality of moneylending: A short history. The Objective
Standard.

Drake, D. (2017, December 6). 2000 Global Crowdfunding Sites to Choose from by 2016. The
Huffington Post.

Hellmann, T., Mospitan, I., & Vulkan, N. (2019, September). Be careful what you ask for:
Fundraising srategies inequity crowdfunding. NBER Working Paper No. w26275

Jackson, D. (2017, July 7). The Netflix prize: How a $1 million contest changed binge-watching
forever. Thrillist.

Rooney, K. (2018, May 31). A blockchain start-up just raised $4 billion without a live product.
CNBC.

Roth, J.D. (2012, November 16). Taking a peek at peer-to-peer lending. Time.

Further Reading
Benkler, Y. (2006). The wealth of networks: How social production transforms markets and
freedom. Yale University Press.
Jemielniak, D., & Przegaliska, A. (2020). Collaborative society. MIT Press.
[Crossref]

Footnotes
1 Some of these companies may also have banking licenses, for example, the UK company Zopa.

 
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_8

8. Value Creation Models and

Competitive Strategy
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Value chain – Value shop – Value network – Coopetition –

Competitive forces

Learning Objectives
After completing this chapter, you should be able to:
– Identify whether an enterprise or business sector is a value
chain, shop, or network, or a combination of these.
– Understand the concepts competition, cooperation, and
coopetition in the digital economy.
– Use Porter’s five forces model in strategic analysis and
planning.

8.1 Introduction
There are several models explaining how value is created within a
company. The classical and perhaps most influential paper concerning
the strategy of industrial companies was written by Michal Porter in
1979 (Porter, 1985). Porter’s model was developed for analyzing the
industrial company’s conversion of raw material into final tangible
products sold to consumers. These companies are categorized as value
chains (► Sect. 8.2) because the production follows a linear chain of
transformations. In 1998, Stabell and Fjeldstad published a paper
proposing two additional types of companies using completely different
ways to produce their product: value shops (► Sect. 8.3) and value
networks (► Sect. 8.4) (Stabell & Fjeldstad, 1998). This chapter
discusses and compares all three value creation models and provides a
broad overview of how the business strategies are different in the three
value creation models.
Most companies in the digital economy are value networks. The
value network is closely related to multisided platforms (MSPs) as
discussed in ► Chap. 10. The following sections briefly review all three
value models to show the differences between them, however, with
emphasis on the value network. The value models presented in this
chapter are supplemental to the production models presented in ►
Chap. 7—production models and value models describe different
aspects of the business operations of a company.
Competition is closely related to value creation. Competitive forces
acting upon a company producing digital goods are considered in ►
Sect. 8.5, while ► Sect. 8.6 discusses different aspects of competition
and cooperation in the digital economy.
As most economic models, these models are only crude but useful
approximations of reality. They are useful in the sense that they identify
some issues that matters to fully understand the roots of value creation
and competition. Or as the British statistician George Box reminds us in
an aphorism attributed to him: “…all models are approximations.
Essentially, all models are wrong, but some are useful. However, the
approximate nature of the model must always be borne in mind” (Box &
Draper, 1987).

8.2 Value Chain

Definition 8.1
The value chain can, in its simplest form, be represented as a linear
chain consisting of three elements: logistics in (retrieval of raw
 material, components, or services), the production of goods, and
logistics out (delivering goods to the market). In addition, there is a
need for common activities such as management, buildings,
inventory, storage facilities, and research.

The value chain is illustrated in ◘ Fig. 8.1. The factory producing

physical goods is a value chain. The actual chain modeling a factory may
contain more elements in a series than the three basic elements in the
definition (Porter included also marketing and promotion and services
in the chain).

Fig. 8.1 The value chain. (Authors’ own figure)

The value chain is just a model by which it is easier to understand

the economy and strategy of the classical industrial organization
producing tangible goods. The value chain has been analyzed in depth
by Michael Porter and other economists.
The logic behind the value chain is that the cost of a single item of a
good can be computed as the sum of the cost of raw materials needed
for each item (m), the cost of producing a single item (p), and the cost of
shipping the item to the market (s). This sum is the direct cost
(marginal cost) of the good. The total cost per item is then the sum of
the direct costs and the cost of the common activities (A) per item. If
the total production is n items, the common activities represent a cost
of A/n per item. The total cost per item is then:
 The major strategy of the chain is to reduce the cost per item. From
the equation, it is obvious that this entails reducing the cost of raw
materials, production, marketing and sales, and the common cost. The
marginal cost (or incremental cost) as n → ∞ is then:

This is the lowest achievable cost per item. Note that this is different
from digital services in which the marginal cost can be zero (MC = 0) as
explained in ► Chap. 6. This emphasizes the difference between the
industrial economy (non-zero marginal cost) and the digital economy
(zero marginal cost).
Reducing the cost of raw materials can be done by bargaining for
lower prices and utilizing the raw materials better. One example is to
increase the yield when producing large semiconductor wafers, thus
making each wafer cheaper. However, the cost of producing large silicon
crystals with few defects is expensive so that there is a balance between
the optimum purity of the crystals and yield. A second example is to
develop mechanical structures requiring less material for the same
structural strength. The cooling towers of nuclear plants are shaped
such that they require less concrete than a cylinder with the same
cooling capacity—their shape just makes the towers cheaper and has
nothing to do with the physical or chemical processes of the reactor. A
third example is to replace existing raw material with new material that
is more easily available. This requires research into new materials, for
example, to produce semiconductors that are more heat-resistant or to
find materials that can replace indium in touchscreens.
The cost of production is reduced by, for example, inventing new
production methods that are faster and more power effective, require
less manpower (e.g., using robots instead of people), or reduce waste.
Maintenance and renewal of machines are also important factors.
Sometimes, better algorithms will do the trick. A German company
producing printed cards for the electronics industry developed an
algorithm for finding the shortest path (also called the “travelling
salesman problem”) for the movements of the robot drilling the holes
into the card, thereby reducing the time it took to produce a single card
(and hence, reducing the price per card).
Out-logistics costs may be reduced, for example, by just-in-time
production, whereby storage requirements and the need for binding
capital over a long period are reduced.

8.3 Value Shop

Definition 8.2
The value shop is a problem-solving organization such as
consultants, health services, engineering companies, and architects.

Value shops earn more money the better and faster they can solve a
problem. Their most important competitive market force is their
reputation. Shops may exist within networks or chains. Some examples
are the advertisement department of newspapers, the consultative
sales department of telecommunications operators, the R&D
departments of the pharmaceutical industry, the design departments of
automobile manufacturers, and the pilot-training centers of airlines.
The shop may, of course, be outsourced to independent companies. For
instance, the newspaper may outsource all its advertisement activities,
including small ads.
Universities and schools are also value shops, as they perform
research, teaching, and innovation. All these activities are problem-
solving tasks; the researcher must find solutions to new challenges, and
the educator must adapt and develop their teaching to target a specific
group of students. Reputation is the most important market force for
both activities.
The value shop is illustrated in ◘ Fig. 8.2. The value shop receives
problems from its customers and solves them using its internal
competence and experience from previous projects. The value shop
may also draw upon external resources to solve complex problems
requiring specific expertise that the value shop does not have. After
solving the problem, a solution is presented to the customer. The value
shop will then evaluate its performance and solution and build
experience. Clever solutions will also contribute to the reputation of the
value shop.

Fig. 8.2 The value shop. (Authors’ own figure)

Value shops depend on customers or clients who have problems

they cannot solve themselves. There are several mechanisms allowing a
value shop to solve problems better or faster than the client himself,
including:
– The shop knows more about the problem than the client.
– The shop possesses specialized methods to deal with the problem.
– The employees are expert professionals in the relevant field.
– The shop can form flexible project teams and is able to reorganize
quickly to solve new problems.
Value shops compete to attract clients with problems to solve.
Competition among value shops depends not only on price but also on
reputation. A lawyer with a good reputation can take higher fees for
consultations and still get more clients than lesser-known or cheaper
lawyers. If the shop’s reputation is destroyed, the company may be out
of business quickly. This happened to Arthur Andersen after the Enron
scandal (Fox, 2002). The company was then accused of accountant
fraud and complicity in hiding large deficits from public view. Though
the Supreme Court acquitted Arthur Andersen for any misconduct, the
company was not able to enter the consultancy market afterward.

8.4 Value Network

Definition 8.3
A value
network
is a business mediating between members of a
market.

The mediation relationships in the value network can be represented as

links between interacting stakeholders. These links form a network
called a value network since the network itself creates value for the
business controlling or operating the network. ◘ Figure 8.3 shows a
general model of a value network. Value networks consist of four
components:
– An organization that provides the services
– Customers or users in the network
– Services enabling interaction between the customers or users
– Contracts that permit customers or users access to the services

Fig. 8.3 The value network. (Authors’ own figure)

Note that the value network is not necessarily a physical network or

a business that owns a physical network. The value network is an
entirely abstract concept. The concept implies that the value network
establishes, for example, an abstract network of selling and buying
relationships between its customers or users. However, there are cases
where the value network also owns a physical network, e.g., a
telecommunications network, but such ownership is not a requirement
for being a value network. Note that in such cases, the mediation
network and the physical network may be two different things. In the
telephone network, the mediating network consists of people calling
one another and not the telecommunications network itself. The
physical telephone network just enables or supports the network of
relationships between the users of the network. The same applies to e-
mail and social media—the telecommunications network is an enabler
for the value network associated with these services. Example of value
network not operating a physical network is the newspaper mediating
between readers and advertisers.
The value network is not a physical network but an abstract
network of relationships between stakeholders.

◘ Figure 8.4 shows four examples of value networks: (1) NY Times,

which mediates between the readers of the newspaper and the
merchants advertising merchandise and services; (2) eBay, which
mediates between the sellers and buyers of goods; (3) Facebook, which
mediates between the users of the social media and the advertisers;
and (4) Kickstarter, which mediates between the project owners and
the funders. Note that these examples are simplified pictures of the
actual business only emphasizing the mediation principles by
suppressing all other business activities taking place in the company.
Fig. 8.4 Examples of value networks. (Authors’ own figure)

Next, let us discuss some particular characteristics of value

networks. The market behavior of some mediation services is different
from most other services. A simple bank facility has two types of
customers: those depositing funds are paid by the bank for being a
customer, while those loaning funds must pay the bank for being a
customer. The price of advertisements depends on the number of
readers of the newspaper—the more readers, the more the newspaper
can charge for each advertisement.
Customers or users may be people or organizations. The users may
be of the same category; for example, friends on Facebook. The users
may also be of different categories; for example, a newspaper is
mediating between readers and advertisers. The latter is an example of
a multisided platform with strong cross-side network effects (see ►
Chap. 10). The same applies to commercial television and social
network services such as Facebook and Twitter. Their revenue from
advertisements is directly related to the number of people using their
services.
As just explained, the competitive strength of several value
networks depends not only on the quality of the product they deliver
but also on the number of users they have been able to capture and, in
particular, how these users stimulate other people to become members
of the network. These are the network effects discussed in ► Chap. 9.
The awareness and popularity of a product is sometimes related to the
diffusion of information of the product via channels other than
advertisement such as word-of-mouth. The “like” button on Facebook
and other social media is another mechanism for increasing the
recognition of a product.
Many of the largest companies in the digital economy base their
businesses on the value network. Examples of value network
companies among the top five companies in the world according to
market cap include Facebook, Amazon, and Google. One of the most
important reasons why these value networks tend to get so big is just
the network effects.
Networks often produce goods that are not stored but consumed
immediately. For instance, it is not possible to store the following for
later use: empty seats in aircraft or trains, surplus energy, unused bits
on the Internet, or empty space in the cargo hold of a truck.
Value networks may offer mutual benefits to its members. This is
the idea behind clubs of different kinds (e.g., literary, musical, bonus
programs, sports)—the more members, the bigger the benefit. The
insurance company can offer better security at a lower price if many
people are using the same insurance company. One reason why they
may lower the price is that the more customers of the same type they
have, the smaller the uncertainty (or variance) of the stochastic product
they are selling (this is called the law of large numbers in statistics).
The cost of one item of a service in a value network can be
expressed as follows:

Here, m is the direct cost (marginal cost) per item, A are the common
costs associated with a product, and n is the number of items produced.
How to measure or compute these three variables is discussed next.

Direct cost (m) In many cases, the direct cost per item is zero (also
referred to as the zero marginal cost property) as discussed in ► Chap.
6. The cost of producing and sending one more bit in the
telecommunications network is negligible. The cost of one additional
passenger on an airplane is almost negligible. The cost of producing
and delivering one copy of an electronic book is negligible. The cost of
adding a new user to social media, such as Twitter or Facebook, is also
negligible. On the other hand, the marginal cost of a physical book is
considerable and includes paper cost, printing cost, storage cost, and
shipment cost. In this case, the printing of the book is done in a value
chain, and the cost is as described in ► Sect. 8.2.

Common costs (A) Since the direct costs are often negligible, the
dominant costs of the value network are the common (or fixed) costs.
These costs are the total cost of running the company and producing
and marketing the services. This is contrary to value chains, in which
the direct cost often is the dominant cost. This also indicates that the
strategy of value networks and value chains are different.

Number of items (n) What is the volume of production? For a

producer of digital books, this is obviously the number of digital books
sold. For an airliner, this may be the average number of passengers per
flight. In a communication network, it may be the number of
subscriptions. It may also be the average traffic carried by the network
(e.g., in terms of the average number of bits sent per unit of time or the
average relative traffic load). Therefore, the number of items may not
be a unique concept in network businesses. To make it even more
complicated, the business may serve a multisided market in which it is
difficult to define what a produced item is, for example, what is the
product produced by Facebook—the number of users; the amount of
information stored about the users; the number of advertisers; and the
number of advertisements? Nevertheless, the formula shows that if the
direct costs are very small, such as for many digital goods and services,
and the fixed costs can be divided on many items, then the marginal
cost is also negligible. This shows that the zero marginal cost property
is common in value networks.

In several value networks, the cost of producing one item of a good

or service is zero.
8.5 Competitive Forces
8.5.1 Porter’s Five Forces
One important strategic evaluation of both tangible and digital markets
is the five forces model proposed by Porter, illustrated in ◘ Fig. 8.5
(Porter, 1979). The purpose of the model is to identify which external
forces that may act on a company and how these forces may influence
the market performance of that company. Based on this knowledge, the
company may develop strategies that counteract the competitive
challenges inflicted by external influence. The model was developed for
strategic analysis of industrial companies that are designed as value
chains. However, the theory can be easily adapted to analyze
competition in the digital economy as shown in the next section. The
five forces are:
– Competitive rivalry is the same as standard competition between
providers of identical goods.
– Threat that new entrants may enter the market making it less
profitable for existing providers of the good.
– Bargaining power of end users and buyers is related to network
externalities such as user feedback and the influence users may have
on the price of goods and the consumption of freemium or free
goods.
– Bargaining power of suppliers has to do with the price they may
charge for raw materials, semi-products, contents, and services
needed in the production of the good. It also includes the availability
of raw materials and other inputs the company needs.
– Threat from substitutes is that the good may be replaced by an
entirely different product offering the same or better user
satisfaction.
► Section 8.5.2 is an adaption of Porter’s theory to digital markets.
The model also includes a sixth force (or rather three such forces) that
also may have strong impact on the market performance of a company.

8.5.2 Porter’s Five Forces Applied to the Digital Economy

Competitive rivalry takes place between companies sharing the same
market. Competition in the digital economy may be on price as in non-
digital markets, but not always, for example, when the price is zero as it
is for several digital goods and services. Competitive rivalry in the
digital economy is complex. In some segments, competition takes place
between companies offering digital services and companies offering
traditional services. This includes e-commerce markets for physical
goods, in which the online shop (e.g., Amazon) may have advantages
since it is accessible anytime from anywhere and offers products that
are not found in the shelves of traditional shops. The advantage of
traditional shops is that the customer can see, touch, taste, and smell
the product.
Competition may take place between companies offering similar
digital services, for example, Facebook and Myspace. In some of these
markets, strong network effects may result in de facto monopolies
where one of the competitors captures the whole or most of the market
(as in the case of Facebook versus Myspace). In other cases, several
competitors may share the market, each having market shares that are
stable over long periods of time, for example, mobile network
operators.
Fig. 8.5 Porter’s five forces. (Authors’ own figure)

Competition may also take place between companies offering entirely

different services to their users. One example is Facebook offering
social networking and Google offering email and web browsing. They
do not compete for users but for money from the advertising business.
This situation may arise in multisided markets (see ► Chap. 10).
Another example is MasterCard serving two markets: cardholders and
merchants. Since the card is accepted almost everywhere, there is no
competition with other credit card companies for attracting new
merchants. The competition is for attracting new cardholders. Airbnb
offer services in two market segments: hosts and guests. Airbnb is
subject to competition in both segments, for example, from hotels and
travel agencies.

New entrants may establish themselves in existing markets. In this

context, the concern is about companies producing the same or
equivalent goods and services. New entrants producing substitutions
are considered below. The general effect of new entrants is that the
profitability of the market for each manufacturer or service provider is
reduced. In cases where investments are high, competition may lead to
the formation of oligopolies, resulting in complex and unstable forms of
competition. The mobile communications markets are oligopolies with
few competitors. In other cases, it may be virtually impossible for new
competitors to enter the market because strong network effects may
have created high lock-in barriers. This is the case for many social
media.

End users and buyers may put additional pressure on the companies,
for example, by creating special interest groups which press for lower
prices or better and more reliable products or for abandoning certain
products altogether (examples from non-digital markets are whale
meat and furs). The buyers may also use new distribution channels.
One example is that more and more people are making purchases over
the Internet, thereby reducing the market for physical shops. Loyalty
programs are sometimes instigated to reduce the bargaining power of
the customers (e.g., bonus programs of airlines).
 Since the use of social media and web browsing creates enormous
amounts of data about the users of the service, the service provider
may be able to extract information about the user that may violate
personal integrity protection laws or are regarded as ethically
unacceptable. This may then cause users to switch to alternative
suppliers or use the service in a way that is less profitable for the
supplier. There have been reactions against Facebook for having
misused customer trust (Confessore, 2018).

Suppliers in value chains deliver raw materials, components, or semi-

finished products. Suppliers in the digital economy may provide
technical support or services, for example, processing and mass
storage, access to the Internet, software, content, or service
supplements. The suppliers are in the position to bargain for prices,
offering different qualities of service to different companies, control
access to facilities and content, and so on. Net neutrality (see ► Chap.
21) delimits the capabilities that the Internet service providers have to
discriminate against different users of the network.

Substitutes are products that may replace other products which offer
the same or an equivalent experience. For example, seafood may
replace meat as nourishment. In the digital economy, substitutes have
become a strong market force. Examples of substitutes include mobile
phones substituting fixed telephone services, and video streaming
replacing broadcast services. The most competitive advantage of
mobile phone manufacturers was originally the design of the radio
modules of the phone, and competition took place between traditional
radio manufacturers. However, as the mobile phones developed into
smartphones (or handheld computers), the competitive advantage
changed to the ability to design complex software that supports the
new functionality, thereby inspiring computer manufacturers to enter
the market. The smartphone then became a substitute for simple
mobile phones produced by a new type of manufacturer. Other
examples of substitutes in the digital economy are e-books substituting
paper books, MP3 players substituting CDs as a medium for the
dissemination of music, and streaming services on smartphone
replacing the MP3 players.
 Later, a sixth force was added to Porter’s model. Different authors
attributed different interpretations of what this force is
(Brandenburger & Nalebuff, 1995):
– Complementors
– Government
– The public
In a strategic analysis of the company, it is recommended to
consider all three alternatives because they will all have an impact on
the company’s strategy.

Complementors are companies that produce or sell products for which

the demand is positively correlated to the demand of a given product.
These products are called complementary products. Complementarity
may be either unilateral or bilateral. Unilateral complementarity occurs
when the product of one company depends on the product from the
other company but not vice versa. Bilateral complementarity means
that each product cannot exist without the other product. Examples of
complementors and complementary products are shown in ◘ Table 8.1.
Table 8.1 Complementors. (Authors’ compilation)

Complementors and Type (unilateral or bilateral)

complementary products
Intel (hardware) and Bilateral: both companies depend on each other to provide a
Microsoft (software) complete service
DVD players and DVDs Bilateral
Printers and ink cartridges Bilateral
Apps and smartphones Unilateral: smartphones offer services that do not involve the
use of apps
Skype and ISPs Unilateral: Skype depends on ISPs for Internet access, while ISPs
do not depend on Skype to offer Internet access

Governments decide rules for competition and oversee that the rules
are followed. In the telecommunications market, regulations may
include license of operation, maximum and minimum price of services
and subscriptions, conditions for lease of network resources, use of the
frequency spectrum, conditions for interconnectivity of customers in
different networks, and number portability. Governments may also
regulate the business of application service providers, for example, via
licensing, taxation, law regulations, and censorship.

The public is made up of more than just users and buyers. Society has
numerous written and unwritten rules that may influence the market of
a company. Some of these rules are ethical; for example, a company may
sell products which are made by suppliers who employ child workers,
cause pollution, or prohibit workers to form labor organizations. This
company may meet strong hostility in the market. Recently, strong
public opposition forced shops to stop selling clothes with certain types
of fur.

Case Study 8.1 Netflix

◘ Table 8.2 shows a simple analysis of the film and video streaming
services of Netflix using Porter’s five forces model plus the three
versions of the sixth force.
Table 8.2 Competitive forces and threats. (Authors’ compilation)

Type Threat
Competition Cinemas, broadcast corporations, Disney +, Amazon Prime, Redbox, Hulu
Plus, and several other providers of equivalent services
New entrants Any new provider of streaming services may capture market shares
from Netflix
Bargaining Because of competition, the bargaining power of uses is mainly on price
power of users
Bargaining Distributors of films may withdraw their contract with Netflix as The
power of Walt Disney Company is doing in favor of its own streaming service
suppliers Disney +. Film studios may also negotiate higher prices and special
conditions for delivery of content. They can also ban Netflix from buying
their content
Substitutes BitTorrent technology (See for example BitTorrent on Wikipedia) based
on P2P distributed file sharing (Popcorn Time was a serious threat to
Netflix before it was discontinued for legal reasons in 2014 (Idland et al.,
2015)). The technology is still a threat to Netflix if intellectual property
rights can be solved in a satisfactory way
Complementors Film studios and video producers for delivery of content and ISPs for
distribution of the streaming service
Government Accusation of tax evasion, high energy usage, and traffic stress on the
broadband networks. Some countries (e.g., Iran and China) block access
to Netflix. Censorship of content
 Type Threat
The public The public may react to the content of films and shows. One problem for
Netflix is that it provides global services and is subject to reactions from
different cultures restricting what is regarded as universally acceptable
content

8.6 Competition, Cooperation, and Coopetition

► Section 4.​3 contains a layered business model for the Internet
consisting of a network layer accommodating the infrastructure
providers and the network service providers (ISPs); a transport layer
(or equipment layer) which is the business arena of the manufacturers
of user equipment and providers of cloud services; and an application
layer containing applications, services, and content. In the layered
Internet model, there is competition between companies within all
three business layers. Companies usually compete within the same
business layer and not between different business layers. At the same
time, companies operating in digital markets cooperate. Companies at
different layers must obviously cooperate to provide digital services to
consumers. For example, an electronic newspaper needs the support of
ISPs to deliver the newspaper to its readers. This is called vertical
cooperation
. Companies in the digital economy may also cooperate at
the same business layer (horizontal cooperation
). Since companies
compete within the same business layer, such companies may both
compete and cooperate at the same time. This is called coopetition
(Shapiro & Varian, 1999)
.

Definition 8.4
Companies compete within the same business layer. The companies
may cooperate both within the same business layer (horizontal
cooperation
) and across business layers (vertical cooperation
).
Coopetition implies that a company simultaneously competes and
cooperates with other companies within the same business layer.

The concepts are exemplified in ► Case Study 8.2.

 Coopetition is particularly important in developing new systems
and services as explained in ► Example 8.1.

Example 8.1 Coopetition and Crowdsourcing: Developing New

Systems or Services
Some of the largest systems or most complex services of the
information and communication technology have been developed
jointly by operators and industries that later have become
competitors offering services to customers, building infrastructures,
or producing user equipment and infrastructure components. This is
an example of coopetition in the digital economy.
The largest effort of this kind is the development of standards for
mobile communications (the 3GPP project). The 3GPP project is
manpowered by participants from more than 700 member
organizations, comprising equipment manufacturers, authorities,
research organizations, and network and service providers. Other
examples are distributed processing (e.g., the CORBA project),
specification of local area networks (e.g., the IEEE Standardization
Association), and cloud computing (e.g., the RECAP project of the EU).
Research groups that are open for participation by anyone are also
established to improve and upgrade existing technologies such as the
Internet, the World Wide Web, and local area network standards.
There are several motives for this type of cooperation:
– The science community prefers that standards should be public and
open for anyone to exploit. In several cases, this is also supported
by governments because open standards enhance competition and
avoid formation of monopolies. For equipment manufacturers, the
potential market for internationally accepted standards is large,
offering good profit margins. For network operators and other
infrastructure providers, the standard opens new opportunities to
expand their businesses to new markets.
– The aim of these projects is usually to define systems and
technologies that can be implemented in a global market. This
implies that industries and researchers from all over the world are
invited to cooperate developing the standard. Experience has also
shown that if the prospects of the project are promising, it is not
difficult to voluntarily man such projects. Moreover, it is simpler to
 attract highly specialized expertise to solve specific problems if the
development process is open for anyone.
– The cost to develop these standards is huge and requires large
resources, in particular, manpower. The estimated cost of
developing a mobile network standard (e.g., 5G) is more than $1
billion and may require several thousand man-years of professional
work. Cooperation reduces the development costs for both
individual manufacturers and operators.
– The solutions these projects end up with usually meet high
performance and technology standards and are better than
solutions developed by a single stakeholder.
This type of cooperative development of standards is an example
of crowdsourcing (see ► Sect. 7.​2.​3). The projects are managed by an
organization (e.g., 3GPP, ETSI, or the World Wide Web Forum), and
the goal is usually defined in vague terms and within an unclear
timeframe, for example, developing the next-generation mobile
system or developing new security algorithms for protecting web
services. The final shape and the details of the end product are
defined as the project progresses, and the end product may be very
different from the product that was initially anticipated. This type of
crowdsourcing takes full advantage of the technological evolution
since the work can change direction as new technological advances
emerge. The best example is the mobile standards which have been
enhanced with new features (e.g., GPRS in GSM and HSPA in 3G) or
become an entirely new standard (e.g., 4G) as soon as the mobile
phone had achieved enough computational power to support these
features.
The participation in these projects is entirely voluntarily, and the
amount of work each participant contributes to the project varies
among the different participants. There are also free-riders who do
not contribute to the development of the project but implement and
make business on the results of the project.
As the complexity of ICT projects have increased, more and more
of the evolution of the technology has been achieved through this type
of cooperative endeavor.
Case Study 8.2 Chat Services and Competition, Cooperation, and
Coopetition
◘ Figure 8.6 illustrates the concepts of competition, cooperation, and
coopetition. The figure shows how a user of a chat service can
choose between various providers at all three business levels
independently. The user may choose among different equipment
supporting chat services (Apple iPhone, Samsung Galaxy, or Google
Pixel), select one of several ISPs (Telenor, Telia, or TDC), and select
one out of four chat applications (Skype, WeChat, WhatsApp, and
Messenger). Each chat application can be used with any combination
of user equipment and ISP. However, chat applications are not
necessarily compatible—if a user selects Skype, they cannot chat
directly with another user using WeChat. This is because there is no
standard among chat applications; therefore, they cannot
interoperate.
 Fig. 8.6 Competition, cooperation, and coopetition in digital markets. (Authors’ own
figure)

On each business level (ASP

, user equipment, and ISP),
companies compete to attract users. Skype competes with WeChat,
WhatsApp, and Messenger; Telenor competes with Telia and TDC;
Apple competes with Samsung and Google. On the other hand, Skype
does not compete with Telenor, Telia, or TDC since they are at
different business layers. This is a truth with modifications since
Telenor is also an ASP offering a chat application through its SMS
service. However, in its role as an ISP, Telenor does not compete with
Skype.
To provide a complete digital good or service, it is necessary for
companies at different business layers to cooperate. For instance,
cooperation between Skype, Telenor, and Apple iPhone is required
to provide the chat service to some users, while Telia, Samsung, and
WeChat must cooperate to provide chat services to other users. This
is vertical cooperation
. These companies may have or not have
formalized cooperation agreements between them. In any case,
through standards, contracts, and agreements, they each provide
elements of the complete digital service.
A Telenor subscriber can communicate with a Telia subscriber, a
TDC subscriber, or a subscriber of any other ISP. International
agreements enforce cooperation between ISPs to ensure
interoperability between operators located in different countries
and within the same country. For the chat service, this implies that a
customer using an iPhone may be a subscriber of Telenor, while the
server providing the chat service is attached to the network of Telia.
It is for the benefit of the market and the users that companies in the
same business layer cooperate. This is called horizontal cooperation
.
These companies also compete for the same customers. As explained
above, this market behavior is called coopetition
.

8.7 Conclusions
Based on the way in which value is created, three types of enterprises
have been defined:
– Value chains transforming raw material into physical products
– Value shops solving problems for their clients
– Value networks mediating between the stakeholders in the market
Many businesses in the digital economy are value networks. One
characteristic of value networks in the digital economy, making them
different from other value creation models, is that the cost of producing
an additional item is zero—or that the marginal cost is zero. This
property allows them to exploit new business models such as freemium
where some basic products are offered for free, while money is charged
for additional features, and multisided markets where some user
groups are not charged for the service they receive, while other user
groups must pay for the services.
The marginal cost of products produced in a value chain is
always larger than zero. The marginal cost of products produced in
several value networks, particularly in the digital economy, is zero.

It has also been shown how the five forces model of Porter—
originally developed for value chains—can be modified to also apply to
value networks. The model can then be used as a valuable tool in
analyzing business strategies for value networks.
The Internet is a layered network consisting of a transport network
transferring information in form of bits, processing equipment
connected to the network, and applications running on the processors.
This structure gives rise to three independent types of stakeholders:
Internet service providers, equipment manufacturers, and application
service providers. There is cooperation between stakeholders at each of
these layers to provide services to the users, and no competition
between stakeholders on different layers. Competition takes place
between stakeholders on the same layer. However, stakeholders at the
same layer may also cooperate, for example, to interconnect users over
multiple networks, to offer complex services consisting of several
components, and to develop and standardize new technologies.

Questions
1. Are the following businesses primarily a value chain, value shop,
or a value network?  
 (a)
Johns Hopkins Hospital  
(b)
AXA  
(c)
Harvard University  
(d)
BMW  
(e)
Apple iPhone  
(f)
Apple App Store  
(g)
Twitter  
2.
Which of the user groups, if any, contribute to the revenues of:  
(a)
The newspaper?  
(b)
The social network service?  
(c)
The stockbroker?  
(d)
Wikipedia?  
3.
How does vertical cooperation between online video and music  
streaming providers and Internet service providers take place?

Answers
1. Value generation model
(a)
 
Value shop  
(b) Value network (insurance company)
  
(c)
Value shop  
(d)
Value chain  
(e)
Value chain (producing iPhones)  
(f)
Value network (mediating between app designers and app  
users)
(g)
Value network  
2.
Revenues  
(a)
Usually both the readers and the advertisers. In many  
cases, the major (and sometimes all) revenue comes from
advertisers.
(b)
Users of the social service do not pay for the usage of the  
service. The revenue comes from other sources, mainly
advertisers.
(c)
Both buyers and sellers of stocks pay a fee to the  
stockbroker.
(d)
Neither readers nor authors contribute with revenues.  
3. The music streaming provider (e.g., Spotify) may have
subscription for broadband access with several network  
providers in different parts of the world to offer streaming
services to millions of simultaneous users. The streaming
service provider may also rent storage facilities in the network
to provide low latency services and reduce the traffic load on the
network. This may become one of the applications of edge
computing in mobile 5G network (for more information, see
Wikipedia article on edge computing)
 Wikipedia article on edge computing).

References
Box, G., & Draper, N. (1987). Empirical model-building and response surfaces. Wiley. The
quotation is found on page 424.

Brandenburger, M., & Nalebuff, B. J. (1995). The right game: Use game theory to shape strategy.
Harvard Business Review, 73(4).

Confessore N. Cambridge Analytica and Facebook. The scandal and the fallout so far. The New
York Times. April 4, 2018.

Fox, L. (2002). Enron: The rise and fall. Wiley.

Idland, E., Øverby, H., & Audestad, J. (2015). Economic markets for video streaming services: A
case study of NetFlix and Popcorn Time. NIKT.

Porter, M. E. How competitive forces shape strategy. Harvard Business Review. March/April
1979.

Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance.

Simon and Schuster.

Shapiro, C., & Varian, H. R. (1999). Information rules: A strategy guide to the network economy.
Harvard Business School Press.

Stabell, C. B., & Fjeldstad, Ø. D. (1998). Configuring value for competitive advantage: On chains,
shops, and networks. Strategic Management Journal, 19, 413–437.
[Crossref]

Further Reading
Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance.
Simon and Schuster.

Stabell, C. B., & Fjeldstad, Ø. D. (1998). Configuring value for competitive advantage: On chains,
shops, and networks. Strategic Management Journal, 19, 413–437.
[Crossref]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_9

9. Network Effects
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Positive feedback – Sarnoff’s law – Metcalf’s law – Odlyzko-

Tilly’s law – Reed’s law

Learning Objectives
After completing this chapter, you should be able to:
– Identify network effects associated with a particular digital
service and the impact the network effects may have on the
temporal evolution of the market.
– Explain how positive network effects may cause slow initial
adaptation of new services and that this may be mitigated by
stimulating early market growth by offering the service for
free initially.
– Demonstrate how the value of networks of different types can
be estimated and use this knowledge in strategic planning.

9.1 Introduction
Definition 9.1 Network Effect
 The network effect (also called network externality or demand-side
economies of scale) is the effect that the number of users or amount
of usage of a service has on the value of that service as perceived
individually by each user (Arthur, 1990).

The network effect is the value a new user adds to existing users in a
network (Shapiro & Varian, 1999). A related term is supply-side
economies of scale, which is the effect the number of users or units
produced has on the costs of production. Supply-side economies of
scale are different from network effects—the former describes the cost
advantages of being large, while the latter des cribes the value of having
many users. This section considers the demand-side economies of scale
—the network effect.
Sometimes a distinction is made between network effect and
network externality. Network externality is then used as the general
term referring to all types of feedback from the market (i.e., negative or
positive), while network effect is only used in the case where this
feedback causes an increase in value of the network (i.e., positive).
Since the nature of positive and negative network effects are essentially
the same, we will use the term network effect referring to both positive
and negative changes in value and, if a distinction is necessary, refer to
them as positive or negative network effects.

Box 9.1 Six Degrees of Separation

Six degrees of separation is the concept that any human being is (at
the most) six intermediaries away from any other human on Earth.
That is, anyone can connect to any other person through a chain of
friends with a maximum of six hops. This is illustrated in ◘ Fig. 9.2.
While by some considered to be an urban myth, research has shown
that the six degrees of separation concept is valid in many social
networks.
The theories of a “shrinking world” were first popularized by the
Hungarian author Frigyes Karinthy in 1929. Karinthy argued that the
modern world at that time was shrinking, primarily because of
recent innovations in communications such as the telegraph, radio,
and telephone. In his paper “The Small World Problem” published in
the journal Psychology Today in 1967, Stanley Milgram showed
experimentally that this was indeed true. He found that the average
social distance between two randomly chosen individuals in the USA
was 5.2 (Milgram, 1967). His observations later became known as
the “six degrees of separation” concept. Stanley Milgram did not use
this term himself; the popularization of the term is attributed to John
Guare and his play “Six Degrees of Separation” from 1990.
Other concepts related to the “six degrees of separation” are the
Erdös number describing the collaboration distance to the
mathematician Paul Erdös based on shared publications (The Erdös
Number Project. Oakland University., n.d.) and the Bacon number
describing the distance to the American actor Kevin Bacon based on
shared movie appearances. The collaboration distances are
amazingly small even between people having published in
completely different fields of mathematics. One of the authors
(Audestad) has a collaboration distance 4 to Erdös and 5 to Einstein!
Facebook has analyzed the average degree of separation between
any two users of the network and found that this distance has
decreased from 5.28 in 2008 to 4.74 in 2011 and 4.57 in February
2016. In the Watts-Strogatz model, which produces random graphs
with small-world properties, the average path length between two
nodes is calculated using the formula lnN/ ln K, in which N is total
number of nodes and K is the average number of links per node. For
Facebook, with 2.2 billion users (nodes) in 2018 and 150 number of
friends (links) per user as suggested by Dunbar’s number (see ► Box
9.2), the average path length is calculated as:ln N/ ln K = ln 2.2 × 109/
ln 150 ≈ 4.29in good agreement with the observed numbers
presented above.

Box 9.2 Dunbar’s Number

Robin Dunbar is a British anthropologist that studied the volume of
the neocortex of various animals and their corresponding social
group sizes. Based on his findings, he predicted the number of people
with whom a human can maintain a stable relationship. His initial
studies suggested a number between 100 and 250, but he later
argued for 150 as a mean group size for communities with high
 incentives to stay together. The latter was based on studies on human
societies, both existing and historical. Dunbar’s number is, then, 150.
It is argued that this is the number of people an individual can call a
“friend,” which translates informally to “people you would not feel
embarrassed about joining uninvited for a drink if you happened to
bump into them in a bar” (Dunbar, 1998).
Dunbar’s number has been applied to social media, business
management, military studies, and workplace organization. The
number has been used to study social networks such as Facebook
and Myspace. The average user on Facebook has 155 friends, which
is close to Dunbar’s number (Knapton, 2016). As quoted by Dunbar:
“The interesting thing is that you can have 1,500 friends, but
when you actually look at traffic on sites, you see people maintain the
same inner circle of around 150 people that we observe in the real
world.” (Knapton, 2016)
Dunbar’s number and his studies of the size of social groups give
empirical input to analytical network models such as Metcalfe’s law.
The communication and formation of social circles might not be that
different in social media as it is in the real world. One way of dividing
people’s social circle is shown in ◘ Fig. 9.3, in which five social
groups are outlined: family and very close friends, close friends,
friends, acquaintances, and strangers.

9.2 Positive and Negative Network Effects

Network effects may be visualized using undirected networks
illustrated in ◘ Fig. 9.1. Network A has 3 nodes and 3 links, network B
has 7 nodes and 11 links, and network C has 11 nodes and 21 links. The
nodes may be individual consumers or users of a specific digital
service, and the links may represent the interaction between the users,
e.g., trading, communication, or any other common interest. Not every
pair of node needs to be connected in these networks. Networks may be
small—as those depicted in ◘ Fig. 9.1—or large such as Facebook, with
more than two billion nodes (users). The number of links is a measure
of the value of a network and is the essential mechanism creating
network effects. Another important concept of networks is the distance
between two arbitrary nodes; that is, the smallest number of links that
must be traversed when travelling from one node to the other. This
distance is important when evaluating how fast innovations diffuse in
networks. The concept is discussed in ► Box 9.1.

Fig. 9.1 Undirected networks. (Authors’ own figure)

Fig. 9.2 Social network. (Authors’ own figure)

Fig. 9.3 Size of human groups. (Authors’ own figure)

Fig. 9.4 Result of positive and negative network effects. (Authors’ own figure)
Fig. 9.5 Positive feedback. (Authors’ own figure)

Fig. 9.6 Adding a new node and links to a network. (Authors’ own figure)
Fig. 9.7 Network effects in social network services. (Authors’ own figure)
Fig. 9.8 Market evolution if all customers are imitators or innovators. (Authors’ own figure)

Fig. 9.9 Facebook users in the period from 2004 to 2015. (Authors’ own figure)

Fig. 9.10 A new user joins the network. (Authors’ own figure)
Fig. 9.11 A broadcast network. (Authors’ own figure)

Fig. 9.12 A network where all nodes connect to each other. (Authors’ own figure)
Fig. 9.13 Numerical examples of Metcalfe’s, Odlyzko-Tilly’s, and Sarnoff’s law. (Authors’ own
figure)

Another concept that has to do with the size of human groups is the
Dunbar number; see ► Box 9.2. The concept is used later in this
chapter to evaluate the strength of certain networks.
We may define two types of network effects: positive network
effects and negative network effects.

Definition 9.2 Positive Network Effect

The network effect is positive if the market feedback causes a
perceived increase in value (e.g., more users)
that stimulates further
increase in value (users).
Facebook is an example where the value of the network increases as the
number of users increases. The positive network effect is caused by
positive feedback from the market.

Definition 9.3 Negative Network Effect

The network effect is negative if the market feedback causes a
perceived decrease in value (e.g., fewer users) that stimulates
further decrease in value (users).

Negative network effects occur, for example, toward the end of life for
multiplayer online games: if the number of players leaving the game
increases, more players are stimulated to leave the game. Another
example is the standard war between VHS and Betamax where positive
network effects in favor of VHS generated negative network effects for
Betamax. This type of negative network effect is also driven by a
positive market feedback (Øverby & Audestad, 2019).
These two cases are illustrated in ◘ Fig. 9.4.
The network effect is also negative if an increased number of users
cause more dissatisfaction with the service, for example, because of
traffic congestion in Internet routers or servers of social media. This
case is a little more complicated than the previous one. If there are
alternative providers offering the same or replaceable services, this
may stimulate the users to switch to another provider. Take the
competition between Netflix and Popcorn Time as example. Because of
congestions in the servers of Netflix in 2014 and first half of 2015,
several users switched to Popcorn Time even though Popcorn Time was
illegal in most countries. This reduced the congestion on the Netflix
servers so that the traffic toward Netflix again increased causing new
congestion. This caused the traffic toward Netflix and Popcorn Time to
oscillate (Idland et al., 2016). This situation ended in mid-2015 when
the Popcorn Time service was taken down in most countries for legal
reasons, and the capacity of the Netflix servers had been improved.
The outcome of this type of negative network effect can be that the
market enters a chaotic or oscillating state where the market
dominance is oscillating between two or more providers. The outcome
may also be that the market reaches a tripping point where it collapses
and is thereafter rebuilt (e.g., the housing bubble in 2008). Note that
oscillating or quasi-stable markets are also subject to positive feedback
from the market.
Negative network effects may be generated by negative reviews and
low ratings of the product, inferior experience expressed by friends,
few active users, or users leaving the product stimulating others to do
the same. Things that may amplify negative networks effects are
insufficient advertisements (invisibility), poor user experience,
technical issues (such as complex login), freeze-out, congestion, long
response times, disturbing differential delays, interruptions, and
frequent downtimes.
Positive network effects may include high ratings, high product
visibility (bandwagon effect), and excellent user satisfaction as
expressed by friends (word of mouth).
Observe again that network effects, whether positive or negative,
are caused by positive feedback
from the market.
Feedback implies that some part of the output from a system is
routed back to the input of the same system and thus causes an effect
on the output. Feedback theory is important in almost all fields of
science: physics, chemistry, technology, social sciences, medicine,
economics, and so on. The first deep analysis of feedback phenomena in
natural systems was done by the American mathematician and
physicist Norbert Wiener (Wiener, 1948). There are two types of
feedback: positive feedback and negative feedback. These are well-
defined technical terms.

Definition 9.4 Positive Feedback

Positive feedback is such that if there is a deviation in the output in
one or the other direction, the feedback will make this deviation
larger. In other words:
– The feedback is positive if more of A (input) produces more of B
(output) and more of B produces more of A and so on until all of A
has been consumed.
– The feedback is also positive if less of A causes less of B and less of
B causes less of A until there is no more A left.
The most important observation is that positive feedback is the cause of
both negative and positive network effects.
Positive feedback may also result in periodic or irregular
oscillations. This is utilized in, for example, high frequency oscillators
producing radio carrier waves for television broadcasting. Applied to
economics, this may be the cause of certain types of business
phenomena such as repeated economic crashes because of
unsustainable growth. One common reason for oscillations in the
market (and in several other cases such as hunter/prey ecosystems) is
long delay in the feedback loop. One example is the “pork cycle” in the
early twentieth century where the demand and supply of pork meat
were out of phase by one year because deficiency (and high price) of
pork meat in one year caused overproduction (and low price) the next
year. Another example is higher education where the delay is several
years causing the labor market to fluctuate between surplus and
shortage of expertise in certain areas: the students choose a particular
education based on the observed prospect of getting a job in that field
at the start of the study. When they finish the education several years
later, the labor market may have changed considerably.

Definition 9.5 Negative Feedback

Negative feedback, on the other hand, reduces the deviations and
tends to bring the output toward a stable equilibrium and keep it
there. If there is an increase in the input A causing an increase in the
output B, then this increase in B will cause a reduction in A bringing
B back close to equilibrium. Similarly, a decrease in A causing a
decrease in B will cause an increase in A again bringing B back close
to equilibrium. The feedback will then counteract any deviation of
the output B away from its equilibrium state. Market stability is one
of the basic assumptions in traditional microeconomic theory where
competition leads to perfect market equilibrium. In this theory, the
market is stabilized by negative feedback loops counteracting any
deviation from equilibrium.

Most markets for digital goods and services are subject to positive or
negative network effects and, hence, subject to positive feedback. This
implies that mainstream economic equilibrium theory does not apply
to such markets.
The points made above are so important (and sometimes
misinterpreted) that it is worthwhile to summarize them. Do not
confuse the terms “positive (negative) network effect” and “positive
(negative) feedback.” Positive and negative network effects are both
driven by positive feedback
from the market. Negative feedback from
the market results in equilibrium markets, while positive feedback
results in non-equilibrium markets. Positive network effects will
drive the market into saturation such as the mobile market. In this
market, everyone sooner or later owns a mobile phone. Negative
network effect will initiate a vicious spiral, in which customers are
leaving the market and the company serving it may face bankruptcy.

◘ Figure 9.5 illustrates how positive feedback stimulates positive

network effects: new users adopting the digital service stimulate other
users to do the same, resulting in a positive network effect that further
stimulates the growth of the service. The final state of most of these
markets is that every user has adopted the service or bought the good
—the market ends up being saturated and cannot increase anymore.
In some markets where competition exists between two or more
supplementary goods (e.g., VHS and Betamax; see ► Chap. 11), one of
the goods may become the good that most customers prefer, generating
a negative network effect on the other product that finally squeezes it
out of the market. Another way of expressing this is that positive
feedback in a competitive market may imply that the strong gets
stronger and the weak gets weaker, also called the Matthew effect, a
term coined by the sociologist Robert K. Merton (Merton, 1968). In this
way, the network effects in competitive markets often produce winner-
take-all markets, where one of the suppliers becomes a de facto
monopoly after all competitors have been squeezed out of the market.
Positive feedback may govern the evolution of a single technology
(search engines), the evolution of a digital service (the World Wide
Web)
, and the competition between technologies or companies
providing digital services (Facebook vs Myspace).
9.3 Characteristics of Network Effects
The network effect may be time-dependent and may even change from
positive to negative as the number of users of the game or service
increases. One example is an interactive video game in which the
gaming experience of the early game attracts more players. When the
number of players increases, the game may become overcrowded,
causing players to leave the game, thereby shifting the network effect
from positive to negative.
The network effect is measured by the amount of interactions or
number of links in the network. This is proportional to the amount of
use of the network resources which in turn is proportional to the time
or amount of attention invested in the network by the users. Remember
that attention is a scarce resource in much demand by the providers
since user attention means opportunities for the providers to sell goods
and services and thereby generate revenues.
The strength of the network effect is thus a direct measure of the
number of links that each new user adds to the network. This is shown
in ◘ Fig. 9.6, in which a new user (node) joining the network connects
to three existing users, thereby adding three new links to the network.
It is not uncommon that a new user connects to all other users in the
network. If a new user connects to exactly one other user in the
network, there are no network effects. This is equal to the Sarnoff type
of network in ► Sect. 9.6.2.
Value networks in general and many digital services have strong
network effects. Examples include Facebook, Twitter, YouTube, Uber,
Airbnb, Skype, MMOGs (e.g., World of Warcraft
), and smartphone app
ecosystems.
Links between users may have different strengths depending on the
importance of the relationship and volume of interaction between the
users. Users may also have different importance in the network
depending on how connected they are. Central users are users with a
high number of links, while marginal users are users with a low number
of links.
A network may also be clustered, meaning that there exist clusters
of networks within the network. One example of this is social media,
where individuals residing within a country are more connected with
each other than with individuals outside their country. Furthermore,
within the same country, people are more connected with each other
within the same town or city. Clusters may also appear among families,
friends, and other socioeconomic groups, for example, workplaces and
people that have similar interest. This leads to networks of networks
with potentially many layers of clusters.
Networks are often modeled as undirected (or symmetrical)
networks. However, the interaction between two users may be
asymmetric—a user sends more information to another user than he
receives in return. Asymmetric users are widespread in digital services
like YouTube and Twitter, where relatively few users are generating
content that is spread to the rest of the network.
◘ Figure 9.7 shows two competing social networking services (A
and B). The figure introduces the terminology of Frank Bass—imitators
and innovators (see ► Chap. 18). The buying habit of imitators is that
they listen to advice from others (i.e., word-of-mouth) and are therefore
likely to buy the version of the product owned by most other people.
The imitator in the figure therefore chooses social network service A.
The innovators do not listen to advice and buy the product of their
liking. The imitators then represent a network effect, in which the value
of the product as perceived by the imitators increases as more people
use the service. The buying habits of innovators are not ruled by
network effects.
Note that the “network” in this context does not refer to physical
networks but to relational networks between users; for example,
between players of the same online game, people and organizations
exchanging emails, family groups, or friends on Facebook. For digital
services, the underlying physical network is, however, a prerequisite for
the existence of the social network. However, sometimes, network
effect may refer to the popularity or ill repute of the physical network.
◘ Figure 9.8 illustrates one of the strategic dilemmas associated
with strong network effects. The figure shows two curves. The S-shaped
curve is the evolution of a market with only imitators, and the second
curve shows the evolution if there are only innovators. As indicated
above, the imitators inflict strong network effects on the market
evolution, while no such effects are present in the case of only
innovators.
 Observe that the initial market growth is extremely slow if all
customers are imitators. It may take several years before the market
has reached a critical mass for which the market penetration is
commercially sustainable (represented by the dotted line). This is the
latency
associated with the market. Moreover, if there are no initial
customers at all, the market never starts to increase—there is no one
there to inspire other customers to buy or use the service. The long
latency is the most difficult strategic dilemma in markets with strong
network effects—the supplier may abandon the product because of the
slow initial growth. On the other hand, the market with only innovators
grows rapidly initially and has low latency. This point is considered in
more detail in ► Chap. 18 where the solution of the Bass equation is
presented.
After a digital service or good with strong network effects reaches
its critical mass, it usually undergoes a period of rapid growth until the
market approaches saturation. The strategic dilemma is how to reach
critical mass in a brief period. If this period is too long, there is a certain
possibility that the service will be prematurely terminated. ◘ Figure 9.9
shows the number of active monthly users on Facebook for the period
from 2004 to 2015 (Roberts, n.d.). Observe that critical mass was
reached in 2008, after 4 years of slow growth. Thereafter, Facebook
grew rapidly. Other social network services, such as Twitter, have
undergone a similar evolution.
Network effects range from concrete and easily identifiable ones to
vague and less obvious occurrences. Easily identifiable network effects
are those where physical objects or users interact directly with each
other. Examples include telephone networks, computer networks,
social media, and multiplayer online gaming. These networks usually
exhibit strong direct network effects, meaning that the network of
users is highly connected.
On the other end of the scale are vague and less obvious network
effects. These include bandwagon effects and tech-performance
network effects. Bandwagon effects include the social pressure to join a
service or buy a good because everyone else is using it. Tech-
performance network effects include the effect that an increased
number of users have on the performance of a technology or service.
Tech-performance network effects are in general easier to identify and
measure compared to bandwagon effects. An example of tech-
performance network effects is peer-to-peer file sharing services such
as BitTorrent. New users joining the BitTorrent network result in, on
average, reduced download times and potentially an increased amount
of available content for other users of the service. Bandwagon and tech-
performance network effects are usually indirect.
Another type of network effects is data network effects, in which
data collected about users or user behavior is used to improve digital
services. Google Search is an example of data network effects since each
search query contributes to refining the Google Search algorithm.
Another example is recommendation systems based on input or
feedback from users.

9.4 Direct and Indirect Network Effects

Network effects may be direct or indirect: direct network effects
take
place when users induce value on other users by the means of direct
interaction between the users; indirect network effects
take place when
the aggregated behavior of the users induces value on other users. For
example, in the case of positive direct network effects, users benefit
from other users in the network because they have more options for
direct interaction. In the case of positive indirect network effects, users
benefit from other users because they, collectively, provide value that is
somehow appreciated by all users in the network without any direct
interaction between the users taking place. Examples of direct network
effects are interactions among members of social media, between
smartphone users, and between buyers and sellers in multisided
markets. Examples of indirect network effects are improvements of
service quality due to user feedback, product reviews, and bandwagon
effects.

9.5 Same-Side and Cross-Side Network Effects

In a network, there may be same-side or cross-side network effects.
Same-side network effects
imply that an increased number of users lead
to an increase in value for other users in the same user group. Examples
are telephones, social networks, and multiplayer online games. These
networks are made up from direct or indirect contact among the users
of the service.
Cross-side network effects
imply that an increase in the number of
users in one user group enhances value in other user groups. These
network effects arise only in multisided markets (see ► Chap. 10). One
example is computers and software—without software, there is no
value in the computer, and without computers, there is nowhere to run
the software. A little less obvious example is smartphones and apps: the
availability of apps increases the value of the smartphone beyond that
of voice and message communications. A final example is third-party
content or service providers in social media. The availability of, for
instance, games on Facebook increases the value for Facebook users as
well as the value for the providers of these games. In this case, there are
positive cross-side network effects.
Note that both same-side and cross-side network effects may be
direct or indirect and positive or negative. Hence, theoretically, there
exist eight different types of network effects—any combination of
positive/negative, direct/indirect, and same-side/cross-side network
effects.
Analysis of network effects associated with digital services includes
identification of all positive, negative, direct, indirect, same-side, and
cross-side network effects and the estimation of the strength of these
effects. A suitable tool for performing such analysis is the stakeholder
relationship model (SRM) presented in ► Chap. 19. The strength of
network effects may be quantified using, for example, Metcalfe’s law.

9.6 Estimating the Value of Networks

9.6.1 Size of Networks
This section provides simple mathematical arguments concerning the
impact that network effects may have on the value of various types of
digital goods or services depending on the way in which users interact.
First, observe that in a population or network of n individuals, there
are:
– n singletons.
– n2 pairs (assuming n >> 1 and counting all one-way interactions
between individuals)
– 2n groups (also assuming n >> 1)
Based on this observation, different laws estimating the value of
digital services are presented next. In the following text, V(n) is a
measure for the total value of a digital service or network, and n is the
total number of users. Value in this context is an abstract concept that
can be a measure of, for example, revenues, market capitalization,
volume of transactions between users, or time spent using a service.
Common representations of value are the number of links between
users in the network and the number of groups a user can be member
of. These are the basis for the calculation of the various network laws.
Note that, in the presented models, V(n) depends only on the number of
users, n. As a general note, the value of a company or a service depends
also on other tangible and intangible assets such as cash, securities,
property, equipment, design value, brand recognition, and
organizational value. However, these variables are not considered in the
following calculations; the focus here is solely on the value that arises
because of the number of users.
The value each new user adds to the network, or the network effect
generated by a single user, is given as:

When a new user joins a network, they link to other users that are
already a part of the network. This is illustrated in ◘ Fig. 9.10.
The tilde notation “~” is used to indicate the growth rate of the value
of a network. For example, V(n)~n indicates that the value of the
network grows (at most) as fast as cn, where c is a constant. This is
equivalent of the “big O notation,” O(n) = cn, commonly used to assess
the growth rate of algorithms in mathematics and computer science.

9.6.2 Sarnoff’s Law

Sarnoff’s law is about the value of broadcast networks such as radio
and television broadcast networks. In such networks, one sender
transmits information to a group of n receivers. Sarnoff claimed that
there is no additional value for new customers to join the network
because others have done so in the past. For the supplier, the value of
the network is the number of customers connected to it, that is:

in which V(n) is the value of the network and n is the number of devices
connected to the network. The value added by a new user to the
network (network effect or feedback term) is FSarnoff(n) = V′Sarnoff(n)~1.
Hence, there is no network effect in this case. Every new user adds only
a single link to the network. This is shown in ◘ Fig. 9.11, in which the
number of customers equals the number of links in the network which,
in turn, equals the total value of the network. The value of a company
providing a broadcast service depends on the number of customers
only since each customer provides a fixed income to the company.
There is no other value created in these networks.
The law is named after David Sarnoff (1891–1971), an American
pioneer of radio and television manufacturing and broadcast. Sarnoff
spent most of his career in the Radio Corporation of America (RCA) and
the National Broadcasting Company (NBC) and was one of the most
influential businessmen in the early days of radio and television.
Sarnoff’s law applies to all kinds of broadcast networks in which
there is no interaction or exchange of value between users or
customers. The only interaction in a broadcast network takes place
between the provider of the service and the users. In addition to radio
and television broadcast networks, there are several examples of other
digital services in which there is no network effect, for example:
– In Google
Search, a user does not benefit from using the search
engine because other people are using it. Therefore, there are no
direct network effects stimulating people to use the search engine so
that, in this respect, Google Search is a Sarnoff network. However,
there may be a weak indirect network effect, hardly recognized by
the users, since search habits of the users contribute to refinement of
the engine’s search algorithm which, in turn, results in more accurate
search results for other users. On the other hand, Google is a
multisided platform, where the users of Google Search generate a
strong cross-side network effect for the advertisement business of
 Google since the number of people using the search engine
determines the fees that Google can charge advertisers.
– Netflix
uses a subscription-based business model. Each subscriber
contributes to the value of Netflix by paying regular subscription
fees. There is no interaction or exchange of value between Netflix
subscribers. On the other hand, Netflix was initially subject to
negative network effects (word-of-mouth) and loss of users to the
illegal Popcorn Time because of overloaded databases (Idland et al.,
2016).
– The value of Wikipedia
depends entirely on the volume and quality
of the articles in the encyclopedia. There is no interaction between
readers, writers, and benefactors and thus no feedback effects
prompting new readers of Wikipedia. The value of Wikipedia for the
reader is the volume and correctness of the content. The value is
independent of who else is reading this content and the number of
writers and benefactors. Wikipedia is a special case where the
network laws do not apply—even Sarnoff’s law—since the value is
independent of the number of users.

9.6.3 Metcalfe’s Law and Odlyzko-Tilly’s Law

Fax machines, telephones, and email accounts are more valuable when
more equipment of the same type is connected to the network. Robert
Metcalfe, one of the inventors of Ethernet, suggested in 1980 that the
value of an Ethernet network would be proportional to the number of
possible transactions between compatible devices connected to the
network, that is:

in which n is the number of devices. The law assumes that all possible
interactions are equally valuable and probable. For small networks,
such as Ethernet, this is a reasonable assumption. This is illustrated in
◘ Fig. 9.12, in which all nodes are connected to each other in a network
consisting of six nodes. The number of links in the network in ◘ Fig.
9.12 is 15 which, according to Metcalf’s law, is a measure for the value
of the network. Note that the number of links in a network with n
nodes, where all nodes are connected to each other, is n(n − 1)/2 ~ n2,
from which Metcalf’s law follows directly.
Later, it was argued that Metcalfe’s law is also valid for social media
network, such as Facebook, Twitter, and other digital services in which
the most important feature is interactions between individuals. A
recent paper (Zhang et al., 2015) indicates that Facebook follows
Metcalfe’s law quite accurately and that the value can be calculated as
V(n) = 5.7 × 10−9 × n2 ~ n2. However, this conclusion depends on how
value is defined and how the number of users is counted.
The network effect (or feedback term) of Metcalfe’s law is
FMetcalfe(n)~n. This is the most commonly used value for network effects
in economic research. The value was, in fact, used long before Metcalfe
suggested it in 1980. This feedback value is the basis for the diffusion
model of Frank Bass from 1969 (see ► Chap. 18).
Andrew Odlyzko and his coworkers have argued that Metcalfe’s law
is incorrect—in particular, concerning social media networks—since
Metcalfe’s law assumes that all possible transactions have equal value.
In a large population, Metcalfe’s law gives an overestimation of the
value of the network since each individual will only interact with a
small number of other individuals, and not all interactions between
them will be equally strong. Based on this argument, Odlyzko and Tilly
proposed the alternative law:

This law is called Odlyzko-Tilly’s law. The network effect is now

reduced to FO − T(n) ~ ln n.

Box 9.4 Derivation of Odlyzko-Tilly’s Law

Odlyzko-Tilly’s law is derived applying Zipf’s principle to the
frequency of interactions an individual has with other individuals.
Zipf’s principle is an empirical law based on the observation that
several sequences in nature and society (e.g., frequency of words,
size of cities, and length of rivers) follow a rank distribution (called a
“Pareto distribution” in statistics) in which the most frequent or
largest item is twice as frequent or large as the second item in the
sequence, three times as frequent or large as the third item, and so
on.
By applying this ranking principle on interactions between
people, the total number of transactions T between one individual
and all other individuals is:

Applying Zipf’s ranking principle, the assumption is that an

individual has c transactions with the person who is closest to him,
only half as many with the next closest person, one-third as many
with the third closest, and so on. The value of the network is then VO −
T(n) = nT~n ln n.
Another way to derive Odlyzko-Tilly’s law is based on the
connectivity of random graphs. First, assume that individuals are
nodes in a random graph and that the interactions between the
individuals are the links in the graph; that is, two individuals
interacting with each other are connected by a link, while there is no
link between two individuals who are not interacting with each other.
The simplest random graph is the Erdös-Rényi (ER) graph. In ER
graphs, the probability, p, that a link exists between any two nodes is
the same for all pairs of nodes in the network.
It can be shown that the ER graph becomes connected if p = c ln
n/n for some constant c. This threshold is sharp. If the link
probability is increasing slightly slower with increasing n, for
example, p = c ln n/na, a > 1 + ε and ε is an arbitrarily small number,
large parts of the graph will be unconnected. If the probability is
increasing slightly faster with increasing n, for example, p = c ln n/na,
a < 1 − ε and ε is an arbitrarily small number, the graph will be tightly
connected. It is reasonable to assume that the graph representing
relationships between people is connected—there exists a path from
one person to another either directly or via other people. This path is
rather short, as revealed by observations made by Milgram, leading
to his law of six degrees of separation; that is, the distance between
people is seldom more than six links, in which a link is from one
person to another person personally know. Since each of us has few
direct links to other people, it is reasonable to assume that the graph
 is lightly connected so that p~ ln n/n is a good approximation of the
link probability of the relationship graph between people. Since there
are n2 possible links, the total number of links N and the value of the
network V(n)~N is VO − T(n)~N~n2 ln n/n = n ln n, and, again, Odlyzko-
Tilly’s law has been derived.

9.6.4 Reed’s Law

In massive multiplayer online games (MMOGs), the players form
groups. The number of possible groups among n players is N~2n
(assuming that n is large). Therefore, it is reasonable to assume that the
value of a digital service in which groups are formed is:

This is Reed’s law (Reed, 2001). Reed is an American computer

scientist and one of the developers of the TCP and UDP protocols.
Reed’s law determines, in general, the value of a network in which
interactions take place in groups. Again, we may use Odlyzko’s
argument that the contribution from large groups is too big and the
actual network effect of group formation is smaller than what is
predicted by Reed’s law. One way to modify Reed’s law is to use
Dunbar’s number, which is the average number of people an individual
knows (see ► Box 9.2). A commonly used value of Dunbar’s number D
is 150.
Let us set the maximum size of a group that can be formed by
people—for example, in an online game—as equal to D (=150). The
number of groups smaller than D that can be formed by n people is:

in which the binominal coefficient is the number of groups of k people

≪ ≪
that can be formed by n people. For large n, n2 nD 2n. The value of
the network is then:
 To see that this is a reasonable assumption, note that by Reed’s law,
the value of the network doubles when a new customer is connected to
the network because VReed(n + 1) ~ 2n + 1 = 2 × 2n = 2 × VReed(n). This
vastly overestimates the value that a single person may have on the
network. In the modified case, the value increases more modestly:

Box 9.5 Derivation of the Modified Law

The formula follows from the observation that:

and that for k < D and D < n/2 (Pascal’s triangle).

Hence,

using the “big O notation” and the fact that D ≪ n.

9.6.5 Summary and Comparison of Network Laws
◘ Table 9.1 summarizes the network laws presented above. These laws
cover a broad array of different networks with very different
underlying value production mechanisms.
Table 9.1 Network laws. (Authors’ compilation)

Law Value of network (V) Network effect (F)

Sarnoff’s law n 1
Odlyzko-Tilly’s law n ln n ln n
Metcalfe’s law n2 n
Law Value of network (V) Network effect (F)
Reed’s law 2n 2n
Modified Reed’s law nD nD − 1

Reed suggests that Sarnoff’s law, Metcalfe’s law, and Reed’s law may
be used to analyze the effect of merging two network companies. The
value of the merged company may then either be proportional to:
– The sum of users of the two companies (Sarnoff), leading to a linear
increase in value
– The increased number of possible interactions between users
enabled by the merger (Metcalfe), leading to a quadratic increase in
value
– The number of groups that can be formed among users in the new
company (Reed), leading to an exponential increase in value
This simple analysis may then, in some cases, uncover otherwise
hidden values and, in other cases, avoid overoptimistic valuations of the
new company.
◘ Figure 9.13 shows numerical examples of Metcalfe’s, Odlyzko-
Tilly’s, and Sarnoff’s law using a linear scale. Observe the significant
differences in value as a function of the number of users for the three
laws. Reed’s law and the modified Reed’s law have not been plotted
since they both increase so fast that they are significantly steeper than
Metcalfe’s law and would, therefore, almost overlap with the vertical
axis.

9.7 Conclusions
The key messages in this chapter can be summarized as follows.
– In some markets, the temporal evolution of the market is stimulated
by positive feedback from the market. This is what is called network
effects. The most important source causing feedback is the number
of customers joining or leaving the market.
– The network effect is positive if an increase in the number of
customers causes further increase in the number of customers.
– The network effect is negative if customers leaving the market
stimulate other users to do the same. The network effect is also
negative if an increase in the number of customers reduces the
quality of the service, stimulating users to leave the market.
– The market may oscillate or become chaotic if there is significant
delay from cause to action.
– Both negative and positive network effects are caused by positive
feedback from the network. Negative feedback stabilizes the market
at a nominal level; that is, counteracting any fluctuations away from
stability.
Network effects are common in the digital economy and are, as
such, important for strategic business analysis in several digital
markets. An important observation is that markets with strong positive
network effects grow very slowly initially and may be abandoned
prematurely.
The network effects depend on the interactions between the
customers. There are three basic types of markets:
– There is no interaction between customers—leading to Sarnoff’s law.
– The customers interact in pairs—leading to Metcalf’s and Odlyzko-
Tilly’s laws.
– The customers interact in groups—leading to Reed’s law.

Questions
1.
Assume that Network A in ◘ Fig. 9.1 grows to Network B and  
finally to Network C. Assume that each link has a value equal to
1. Applying the definition of network effects, how does the value
per user increase as this network evolves?
2.
Snapchat is a social media service available for smartphones.  
(a)
Identify network effects in Snapchat (positive, negative,  
direct, indirect, cross-side, and same-side network effects).
(b)
How are network effects in Snapchat different from  
network effects in Facebook?
3. Does Sarnoff’s law give an accurate valuation of Instagram?

 
4.
Draw two graphs—one linear and one logarithmic—that show  
the network effect as a function of the number of users using (1)
Sarnoff’s law, (2) Metcalfe’s law, and (3) Odlyzko-Tilly’s law.
How would you interpret the graphs? (The network effect is
defined in ► Sect. 9.6.1.)
5.
Assume two competing social media companies, A and B, with  
NA = 10 million users and NB = 20 million users. Assume that
these two companies plan to merge, making a new company, C,
with NC = NA + NB = 30 million users. Assume further that the
valuation of the companies depends on the number of users
only. What is the total gain (in percentage) from such a merger
when estimating the company’s value using:
(a)
Sarnoff’s law?  
(b)
Metcalfe’s law?  
(c)
Odlyzko-Tilly’s law?  
6. Consider a social media network that you are a part of (e.g.,
Facebook, Twitter, or Snapchat).  
(a)
How many users are there in total in the chosen social  
media network?
(b)
How many users do you link to in the chosen social media  
network?
(c)
How will you rank and quantify the importance of your  
links?
(d)
Does Metcalfe’s law or Odlyzko-Tilly’s law best describe  
 y y
the value that the social media network offers to you?
7.
 
Facebook has, by 2018, approximately 2.2 billion monthly active
users.
(a)
Based on this number, calculate the total value of Facebook.  
You can use the formula V(n) = 5.7 × 10−9 × n2.
(b)
What does this value represent?  
(c)
How many users does Facebook need to have for a total  
value of 1?

8.
Which of the three laws (Sarnoff, Metcalfe, and Reed) are valid  
for the valuation of the following digital services: Google Search,
Gmail, World of Warcraft
, Wikipedia, Netflix, Twitter, YouTube,
eBay, and Facebook?

Answers
1.
Assuming that the value of the network is the number of links  
per user, we find:
1.
Network A: 3/3 = 1.  
2.
Network B: 11/7 = 1.57.  
3.
Network C: 21/11 = 1.9.  
The network effect is positive since the value increases
as the number of users increases.
2. Snapchat is used for photo sharing, videos chat, and instant
messaging.  
 1. Snapchat is subject to positive network effect (the number
of users). This is a direct same-side network effect. Snapchat  
also offers advertisements. There is then a direct cross-side
network effect from the number of users to the advertisers.
2.
Facebook (more than 2.2 billion users) is much larger than  
Snapchat (about 230 million users). In accordance with
Metcalf’s and Odlyzko-Tilly’s laws, the network effects of
Snapchat are much weaker than for Facebook.
3.
Instagram is a value network mediating between users sharing  
photos and videos and is not a broadcast network. Therefore,
Sarnoff’s law does not apply to Instagram.
4.
The strength of the feedback is defined as the value each user  
adds to the network. The feedback function is shown in ◘ Fig.
9.13 for the linear case.
– In Sarnoff’s law, the value that each user adds to the network
is 1; that is, there is no feedback in this case.
– In Odlyzko-Tilly’s law, the additional value each new user
adds to the network increases as lnn as the number of users
increases. For a large network such as Facebook, the user
adds approximately 20 units of value to the network.
– In Metcalf’s law, the additional value that each user adds to
the network is n. For Facebook, this is approximately 2 billion,
that is, 100 million times more than Odlyzko-Tilly’s law.
5.
The gain is defined as:  

The result is shown in the table.

Value model Company A Company B Company C Gain in percent

10,000 users 20,000 users 30,000 users
Value model Company A Company B Company C Gain in percent
10,000 users 20,000 users 30,000 users
Sarnoff 10,000 20,000 30,000 0
Metcalf 1 × 1014 4 × 1014 9 × 1014 80

Odlyzko-tilly 1.6 × 108 3.4 × 108 5.2 × 108 4

6.
Look for information on Wikipedia or the web for your choice of  
social media network.
7.
Value of Facebook.  

(a)
This value can be a measure for (i.e., proportional to) the  
number of user interactions, revenues, or stock market
value.
(b)
To have value 1, the number of users must be:  

8. Value model:
1.
 
Google Search: Sarnoff (no interaction between users).  
2.
Gmail: Sarnoff (there are other email services).  
3.
World of Warcraft
: Reed (formation of groups of players).  
4.
Wikipedia: none of them—the value of Wikipedia is  
independent of the number of users.
5.
Netflix: Sarnoff (no interaction between users).  
 6. Twitter: Metcalf (depends on the number of followers).
 
7.
YouTube: Metcalf (number of viewers depends on the  
number of videos producers).
8.
eBay: Metcalf (depends on the size of the marketplace).  
9.
Facebook: Metcalf (depends on the number of friends).  

References
Arthur, W.B. (1990). Positive feedbacks in the economy. Scientific American, No. 262.

Dunbar, R. (1998). Grooming, gossip, and the evolution of language. Harvard University Press.

Idland, E., Øverby, H., & Audestad, J. (2016, November 23–25). Economic Markets for Video
Streaming Services: A Case Study of Netflix and Popcorn Time. NIK 2016, Ålesund, Norway.

Knapton, S. (2016, January 20). Facebook users have 155 friends – But would trust just four in a
crisis. The Telegraph.

Merton, R. K. (1968, January 5). The Matthew Effect in Science. Science, 159.

Milgram, S. (1967). The small world problem. Psychology Today, No. 2.

Odlyzko, A., & Briscoe, B. (2006, July 1). Metcalfe’s law is wrong. IEEE Spectrum: Technology,
Engineering, and Science News.

Øverby, H., & Audestad, J. A. (2019) Temporal market evolution of interactive games. SSRN.
Available at https://​ssrn.​com/​abstract=​3412902

Reed, D. P. (2001, February). The law of the pack. Harvard Business Review.

Roberts, P. (n.d.). The most important Facebook statistics for 2017. Our Social Times, Cambridge.

Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to the network economy.
Harvard Business School Press.

The Erdös Number Project. Oakland University.

Wiener, N. (1948). Cybernetics: Or control and communication in the animal and the machine.
MIT Press.

Zhang, X.-Z., Liu, J.-J., & Xu, Z.-W. (2015). Tencent Facebook Data Validate Metcalfe’s Law. Journal
of Computer Science and Technology, 30(2).

Further Reading
Dorogovtsev, S. N., & Mendes, J. F. F. (2003). Evolution of networks: From biological nets to the
internet and WWW. Oxford University Press.
[Crossref]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_10

10. Multisided Platforms
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Two-sided platform – Multisided platform – Digital MSP –

Tangible MSP – Reseller

Learning Objectives
After completing this chapter, you should be able to:
– Identify the different user groups of a multisided platform.
– Identify the value proposition for each user group and the
combined effect on the aggregate business offered by the
platform.
– Analyze same-side and cross-side network effects governing
the evolution of the different services offered by the platform.

10.1 Introduction
In a multisided platform (MSP), two or more distinct user groups
interact to produce mutual benefits for each other. In many practical
cases, there are just two groups, in which case, the multisided platform
becomes a two-sided platform. There is no essential difference between
two-sided and multisided platforms. The most important difference is
that the multisided platform consists of many more business roles and
interactions between these roles. This chapter will therefore mainly
discuss two-sided platforms, keeping in mind that the theory is valid
also for multisided platforms. For simplicity, we will also refer to two-
sided platform as MSP.

Definition 10.1 Multisided Platform

A multisided platform (MSP) enables direct interactions between
two or more distinct user groups in which all user groups are
affiliated with the MSP (Hagiu & Wright, 2015).

The MSP is one of the most profitable innovations of the digital

economy. Some of the largest companies in the digital economy base
their business models on an MSP, including Google, Facebook, and eBay.
Some MSPs have even become market leaders in their industry since
there is tremendous value in connecting different user groups. Still,
there seems to be a considerable potential for new MSPs in several
business and industry sectors, as well as in the public domain.
A comprehensive overview of the literature on the economics of
multisided platforms is presented by Sanchez-Carlas and Leon
(Sanchez-Cartas & Leon, 2018). The overview shows that the general
approach to dynamic behavior of platforms is based on standard supply
and demand theory and agent-based behavior theory. These methods
provide insight into important areas such as pricing and price
dependence between the different markets supported by the platform,
the origin of network effects, coupling between platform services,
behavior of users, competition between platforms offering equivalent
services, problems associated with regulations, and formation of de
facto monopolies.
In this chapter, the focus is on the temporal evolution of multisided
platforms triggered by strong network effects and not on pricing and
competition in equilibrium markets. Several of these markets are
usually not in equilibrium because they have not reached saturation
where growth has stagnated (e.g., most social media platforms).
Moreover, some of the platforms (e.g., Facebook) have become de facto
monopolies. In these cases, standard supply/demand theories do not
apply.
10.2 Characteristics of Multisided Platforms
MSPs offer several benefits to the users, including:
– They offer uniform, simple, and secure login procedures and ease of
use.
– They offer simple and transparent ways of connecting customers
with matching interests through built-in matchmaking mechanisms,
for example, matching suppliers and consumers in a trade process or
providing information to marketers to support targeted and precise
advertisement campaigns.
– They offer efficient transaction management, thereby simplifying
search, reducing management costs, and streamlining administrative
and payment processes.
– They may also assist in trust-building between users within the same
group or between users in different groups.
The MSP is related to the concept of value networks described in ►
Chap. 8 since the value network is an organization or company offering
mediation services either within a single group or between different
groups, in which case, the value network is equivalent to an MSP. ►
Chapter 8 explains the strategy of value networks in general. In this
chapter, the attention is on the mediation activity itself and the techno-
economic platform supporting it.
It is assumed that the mediation between the user groups is
asymmetric. Asymmetry means that the platform is serving two or
more user groups with different motives for using the platform
services. ◘ Table 10.1 shows examples of MSPs.
Table 10.1 Examples of MSPs. (Authors’ compilation)

MSP Type of business User groups Platform type

New York Times Newspaper Readers and advertisers Digital
eBay Electronic marketplace Sellers and buyers Tangible
Facebook Social networking service Users and advertisers Digital
Uber Sharing service Drivers and passengers Tangible
Airbnb Sharing service Hosts and guests Tangible
MasterCard Point-of-sale transactions Merchants and cardholders Digital
MSP Type of business User groups Platform type
Kickstarter Crowdfunding Borrowers and investors Digital

There are two main types of MSPs: digital MSPs and tangible MSPs.
Digital MSPs mediate the exchange of digital goods and services, while
tangible MSPs mediate the exchange of physical goods and non-digital
services. Facebook and MasterCard are examples of digital MSPs,
whereas Uber and Airbnb are examples of tangible MSPs. Tangible
MSPs have also been termed online-to-offline (O2O) MSPs (McAfee &
Brynjolfsson, 2017). ◘ Table 10.1 lists some two-sided platform
businesses.
Some of the major characteristics of MSPs are associated with
network effects, pricing, competition, business ecosystem, and market
regulation (Ardolino et al., 2020).

10.2.1 Network Effects

The MSP may be subject to complex network effects not only involving
each user group but also between user groups. Network effects are so
important for MSPs that they are discussed in more detail in ► Sect.
10.3. The network effects may sometimes lead to winner-take-all
markets where one platform supplier after some time becomes a de
facto monopoly.

10.2.2 Pricing
The pricing model and, hence, the way revenues are generated may be
complex. For example, one user group are offered platform services for
free, while other customer groups must pay for the services they
receive (e.g., Facebook and Google Search); customers may pay per
interaction or amount of service they receive, or they may pay a fixed
monthly subscription fee (e.g., eBay and Airbnb); customers may pay
for some types of services and receive other services free of charge
(electronic newspaper); or any combination thereof.
It is a common feature of MSPs that some user groups subsidize the
other user groups by levying differentiated charges among the user
groups. Sometimes, the MSP gets all its revenues from only one user
group while providing services for free to other user groups.
 An MSP often benefits from a reduced price of the goods or services
mediated between the sellers and the buyers. This is because low prices
mean more sales and, potentially, more buyers of the goods, increasing
the usage of the platform and, in turn, add value to the platform in
terms of increased cross-side network effects. In this respect, MSPs
have incentives to reduce prices in the business areas in which they
operate.
In several of these pricing regimes, supply/demand theories are
meaningless since the two variables are decoupled in market segments
where the good is traded for free.

10.2.3 Competition
Three basic types of competition are identified for MSPs. The platform
may compete directly with other platforms offering the same services
(e.g., between Facebook and Myspace and between electronic
newspapers) or compete with entirely different platforms for certain
types of customers. Facebook and Google compete for attracting
advertisers to their platform even though the two platforms belong to
completely different business segments. This type of competition may
seem counterintuitive but is the most important competitive challenge
for several platform operators. Finally, there may also be competition
between the customers of the same type (e.g., between drivers offering
services over the Uber platform). This makes competition in the MSP
business more complex than in most other businesses.
One particularly interesting group competing in the advertisement
business is the influencers promoting various products on their blogs
over social media platforms (e.g., Facebook, Instagram, and YouTube).
The influencers run two-sided platforms (readers and advertiser) on
the Internet without owning any kind of infrastructure except their
own smartphones, tablets, cameras, or personal computers.

10.2.4 Business Ecosystem

Because of competition as just explained, the ecosystem for MSPs is
more complex than other businesses. For this reason, the MSP must
sometimes include other businesses in their ecosystem analysis that
are seemingly unrelated to the primary business areas of the platform
making the overall business model more complex and unpredictable.
Because of the complex interactions among the various types of
customers, the use of standard business modeling tools may not
capture all strategic issues related to MSPs. Business models often fail
to take the effects of interactions between user groups (e.g., cross-side
network effects) into account but rather treat them as independent
business segments. One way of characterizing the MSP in terms of its
ecosystem is considered in ► Sect. 10.4.

10.2.5 Market Regulations

The mere existence of two or more user groups and sometimes also
strong network effects both within the same user group and between
the various groups make it difficult to find one regulatory regime that
supports fair competition and avoids formation of monopolies.
Facebook is a de facto monopoly in the segment of social media services
but not in the advertisement segment. Moreover, intricate competition
and complex ecosystem make it even harder to regulate MSP
businesses. The complexity itself may make it difficult to identify what
needs to be regulated, what can in fact be regulated, what are the actual
effects of regulation, and how can lock-in be avoided.

10.3 Network Effects and MSPs

Network effects (see ► Chap. 9 for a general discussion of network
effects) play an essential role in MSPs as illustrated in ◘ Fig. 10.1 for a
two-sided platform. Here, there are two distinct user groups. The
platform mediates between these two user groups as well as between
users within each user group. If the network effect involves only one
side, it is called a same-side network effect. If it has an impact on the
other side, it is called a cross-side network effect. Positive, negative,
direct, and indirect network effects may be present within user groups
and between user groups. The dynamics of an MSP may be complex
since there are dependencies and network effects within user groups
and between user groups.
Fig. 10.1 Network effects in a two-sided platform. (Authors’ own figure)

It is worthwhile to repeat that network effects are positive

feedbacks from the market and that the mere nature of positive
feedback is such that the market cannot be in equilibrium unless the
market is saturated (there are no more potential customers). Therefore,
equilibrium theory is not applicable to markets where network effects
are present.
The Bass equation (see ► Chap. 18) is an example of how the
market evolves if there are same-side network effects present. The
outcome a simple cross-side network effect has on the temporal market
evolution is analyzed in ► Box 10.1 using a simple mathematical
model.
The simple model in ► Box 10.1 shows that the network effects
may lead to slow service adoption by one (or even both) user groups in
two-sided markets. Strategically, this may deceive the platform provider
to believe that this user group will not adopt the service and, therefore,
prematurely abandon it. This is a general problem associated with
positive network effects as explained in ► Chap. 8 and shown
mathematically in ► Chap. 18. This problem is particularly vital if the
revenues of the platform provider depend mainly on the service of the
customer group subject to the positive network effects. The revenues of
Facebook depend almost entirely on the number of advertisers they can
capture, which, in turn, depends strongly on the number or users of the
social platform. For this reason, the revenues of Facebook increased
slowly initially as shown in ► Case Study 10.1.
The networks effects of MSPs consisting of ordinary users and
advertisers (e.g., Facebook) can then be summarized as follows. The
number of ordinary users enforces a strong positive network effect on
the number of advertisers. The number of advertises, on the other
hand, will have little or no effect on the number of ordinary users so
that the there is only cross-side network effects in the direction from
users to advertisers. There are no network effects among advertisers
though some of them may be competitors. Among the ordinary users,
there are strong positive network effects; this is a major cause of
growth of this type of MSPs.
In the electronic marketplace consisting of sellers and buyers (e.g.,
eBay), the sellers prefer many buyers, and the buyers prefer many
sellers. Therefore, this platform has positive cross-side network effects.
On the other hand, the existence of many sellers means increased
competition among seller, possibly generating a weak negative same-
side network effect. There is generally no same-side positive network
effect concerning buyers; however, weak network effects may be
present generated by user reviews and word of mouth.

Box 10.1 Dynamics of Markets with Cross-Side Network Effects

A simple mathematical model for the temporal evolution of a two-
sided platform with a single cross-side network effect is shown in ◘
Fig. 10.2. This simple model can easily be extended to models with
multiple same-side and cross-side network effects and more than
two types of customers.
Fig. 10.2 Two-sided platform with cross-side network effects. (Authors’ own figure)

There are two types of customers, A and B. Initially there are N

potential customers of type A and M potential customers of type B.
Customers of type A adopts the platform service offered to them at a
fixed rate p, symbolized by a valve controlled by the parameter p. The
adoption rate of customers of type B is proportional to the number of
customers of type A, i.e., the adoption rate of type-B customers is qA.
This is a cross-side network effect from customer-side A to customer-
side B. The flow parameters p and q are assumed to be constants;
otherwise, the equations cannot be solved analytically. The flow
parameters may depend on other factors such as price, service
promotion, and visibility. This is not included in this simple model
where the aim is to show how the growth in one customer segment
may influence the growth in the other customer segment.
The flow rate of customers is equal to the number of users
adopting the service per unit time. This is, by definition, equal to the
time derivative of the number of customers having adopted the
service at given time. Hence, the flow of customers of type A = dA/dt
= p(N − A) and the flow of customers of type B = dB/dt = qA(M − B),
where N − A and M − B are potential customers who have not
adopted the services yet. This gives the following set of coupled first-
order differential equations for the evolution of the two markets:
 The first equation is solved immediately giving:

Inserting this in the second equation gives:

with solution:

For small t, B increases as:

which is much slower than linear increase for small values of t.

The evolution of the relative market size (A/N and B/M) is shown
in ◘ Fig. 10.3. The abscissa is the time in years, the ordinate is the
relative number of customers, and the flow parameters in the
example are p = 0.17 and qN = 0.21.
Fig. 10.3 Temporal market evolution for a two-sided platform. The parameters used are p =
0.17 and qN = 0.21. (Authors’ own figure). The abscissa is the time in years

Because of the feedback from customer-side A to customer-side

B, the growth of type-B customers will follow an S-curve and lagging
the growth curve for type-A customers. In the example, it takes about
6 months until 10% of the market of type-A customers has been
captured. At that time, only about 0.5% of the potential market of
type-B customers has been captured. It takes 2 years to capture 10%
of the market of type-B customers. If the revenues depend only on
the number of type-B customers, the cross-side network effect may
result in too small revenues to sustain the service initially. The
service may then be prematurely terminated.

Case Study 10.1 Growth of Facebook

The theory in ► Box 10.1 can be used to study the temporal
evolution of Facebook. ◘ Figure 10.4 shows the evolution of users
and revenue of Facebook from 2008 to 2019 (Facebook’s annual
revenue from 2009 to 2019, 2019; Number of monthly active
Facebook users worldwide as of 4th quarter 2019, 2020). The
revenues are assumed to be proportional to the potential number of
advertisers. The theoretical curves are calculated using the formulas
of ► Box 10.1 with parameters N = 7.8 × 109 (the current world
population), M = 1.9 × 108 (the current worldwide number of
companies—as estimated by ► datapo.​com), p = 3.7 × 10−2, qN = 1.9
× 10−2, and the average revenue generated per advertiser is 10 USD.
The similarity between theoretical and observed curves is striking.

Fig. 10.4 Market and revenue evolution of Facebook. (Authors’ own figure)
10.4 Business Ecosystem and MSPs
One generic model for MSPs consists of six types of stakeholders in
addition to the platform itself: content providers, advertisers,
developers, professionals, merchants, and consumers (Gautier &
Lemesch, 2020). A platform may offer services to all six categories or to
only some of them. Similarly, several stakeholders belonging to the
same category receiving similar or different services from the platform
may be connected to the platform so that the overall ecosystem of the
MSP may be overly complex. The configuration is shown in ◘ Fig. 10.5.

Fig. 10.5 Generic ecosystem for MSPs. (Authors’ own figure)

Content providers may use the platform for streaming of media

(Spotify), offering specialized content to certain user groups (e.g., via
Facebook), production and publishing of digital content (e.g., Google’s
YouTube and Amazon’s Kindle), and support of video games
(Facebook).
Advertisers use the platform for marketing, promotion, campaigns,
publicity, and targeted and personalized advertisements, usually based
on information the platform has gathered about each individual
customer. This is a primary business of, for example, Facebook and
Google.
Developers use the platform to develop new services or goods, most
notably, apps for Apple’s App Store and Google’s Play Store.
Professionals are companies or organization receiving services that
are used in their own production of digital or physical goods, for
example, cloud computing, production of statistics (e.g., based on user
data from Facebook, Google, or Amazon), online accounting support,
collaboration services (e.g., iWork from Apple), document sharing,
backup and security services (e.g., WorkDocs from Amazon), and
physical or digital delivery services.
Merchants use the platform for online shopping and payment of
both physical and digital goods.
Consumers are the end users of digital services using smartphones,
tablets, and other computers to access and process the information, or
they are the receivers of physical goods ordered over a digital platform.

10.5 Trade Process of MSPs and Resellers

The MSP is one of several ways to organize a company. Another
common type of organization is the reseller illustrated in ◘ Fig. 10.6.
The reseller purchases goods and services from a producer and resells
these goods and services to the buyer. The buyer receives the goods and
services from the reseller, who, in turn, receives payment from the
buyer.
Fig. 10.6 The reseller. (Authors’ own figure)

One important requirement for being an MSP is that the platform

enables a direct contact between the user groups attached to it. An
electronic marketplace, such as eBay, is different from a supermarket in
this respect. The supermarket is not an MSP since there is no direct
contact between the shopper and the supplier. The supermarket is a
reseller. On the other hand, eBay may sell the same goods as the
supermarket, but eBay is an MSP since there is direct contact between
supplier and buyer. For similar reasons, a taxi company is not an MSP,
while Uber is.
An example of a reseller is Amazon: Amazon buys books from
publishers and resells them to readers in the same way as any other
bookstore. Resellers of telecommunications services are another
example of resellers—they purchase bulk traffic capacity from a
network operator and resell it in subscription packages to the
customers.
The MSP platform may be quite simple. It may consist of a single
app loaded down from the Internet and processed on the smartphones
of the supplier and the customer. In other cases, the platform may be
complex such as the platforms of Google and Facebook. The MSPs may
then be companies ranging in size from one-person enterprises to large
corporations. Most sharing economies are two-sided or multisided
platforms.
◘ Figure 10.7 illustrates a trade process for an MSP. The MSP
mediates contact between the seller and the buyer. Observe the
similarities between the MSP and the value network in ► Chap. 8. The
trade—including the flow of goods, services, and payment—is done
directly between the seller and the buyer and facilitated by the MSP.
The MSP is not directly involved in the exchange of goods.

Fig. 10.7 The trade process of multisided platforms. (Authors’ own figure)

10.6 Conclusions
The multisided platform (MSP) provides services to two or more user
or customer groups at the same time. To be an MSP, there must be
interactions between the groups providing benefits for each user group.
The interactions between the groups are the cross-side network effects.
This gives rise to complex business ecosystems consisting of several
types of stakeholders and the numerous interactions between them.
This makes the analysis of MSPs considerably more complex than
businesses offering services to a homogenous group of customers.
Most businesses in the digital economy are multisided platforms.
Therefore, the understanding of how interactions take place in these
platforms is essential in the analysis of price formation within each
individual user group, how the platform configuration generates
revenues, and how competition takes place in the MSP market.
Competition is particularly challenging since competition may take
place not only between similar platforms but also with platforms in
different business segments, for example, Facebook and Google
compete for attracting advertisers.
Several MSPs offer services for free to some user groups, while
other user groups pay for the services. This flexibility is one reason for
being an MSP. The revenues of the MSP are then determined by the
strength of the cross-side network effect these users impose on the
customers paying for the services. In such cases, standard economic
demand/supply analysis may be meaningless.
Finally, several MSPs are de facto monopolies in one of their
business segments (e.g., as social medium) while competing in other
segments.

Question 10.1 Three-Sided Markets

Postmates is a logistics company mediating between companies
having goods to deliver, engaging couriers for delivering the goods,
and customers. Illustrate Postmates as an MSP and its associated
user groups. Identify same-side and cross-side network effects
within and between user groups. Which of the user groups
contribute to Postmates’ revenues?

Question 10.2 Multisided Platform Laws

Discuss whether Metcalfe’s law and Odlyzko-Tilly’s law can be
applied to multisided platforms.

Question 10.3 YouTube

What is the ecosystem of YouTube? Who generates the revenues?
Identify cross-side and same-side network effects.

Answer 10.1
The configuration of Postmates is shown in the figure. The platform
serves three user groups: companies (restaurants and merchants)
selling the goods, the couriers transporting the good to the
customers, and the customers ordering the good. Referring to ◘ Fig.
10.5, these correspond to merchants, professionals, and customers.
The companies selling the goods contribute with revenue to
Postmates. Postmates pays the couriers for transporting the goods.
The customers pay the company delivering the good, where the price
also covers the cost of engaging couriers.
There are cross-side network effects between all three groups.
The more sellers, the more customers, and the more couriers. The
more customers, the more seller, and the more couriers. There are
no significant same-side network effects (◘ Fig. 10.8).

Fig. 10.8 Network effects in Postmates

Answer 10.2
Yes. In an MSP, there are cross-side network effects. Metcalfe’s law or
Odlyzko-Tilly’s can be applied to calculate the strength of the
network effects or the value that each user group represents. In ►
Box 10.1 and ► Case Study 10.1, Metcalf’s law was used to do this.

Answer 10.3
The stakeholders are uploaders, users receiving free services,
viewers paying for premium services, content providers channeling
 information through YouTube, and advertisers. The revenues are
coming from advertisers and viewers paying for premium services.
Cross-side network effects:
– The more uploaders, the more viewers.
– The more viewers, the more uploaders.
– The more viewers, the more advertisers.
Same-side network effect: uploaders stimulate other uploaders.

References
Ardolino, M., Saccani, N., Adrodegari, F., & Perona, M. (2020). A business model framework to
characterize digital multisided markets. Journal of Open Business: Technology, Markets, and
Complexity, 6, 1–23.

Facebook’s annual revenue from 2009 to 2019. Statista, 2019.

Gautier, A., & Lemesch, J. (2020, January). Mergers in the digital economy. CESifo working paper
No. 8056.

Hagiu, A., & Wright, J. (2015). Multi-sided platforms. International Journal of Industrial
Organization, 43, 162–174.
[Crossref]

McAfee, & Brynjolfsson, E. (2017). Machine, platform, crowd: Harnessing our digital future. W. W.
Norton & Company.

Number of monthly active Facebook users worldwide as of 4th quarter 2019. Statista, 2020.

Sanchez-Cartas, J. M., & Leon, G. (2018). Multi-sided platforms and markets: A literature review.
Universidad Politecnica.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_11

11. Path Dependence
Harald Øverby1   und Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Korrespondenzautor)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Learning Objectives
After completing this chapter, you should be able to:
– Understand the concepts of diminishing returns, increasing
returns, and path dependence in the context of the digital
economy.
– Explain why some economic systems—contrary to standard
microeconomic theory—may end up in one out of several
equilibrium states depending on internal and external forces
acting on the market.
– Analyze how path dependence is generated by positive
feedback from the market or by external forces.

1 Definitions
The general assumption in conventional economic theory is that the
markets are controlled by negative feedback that reduces any deviation
away from market equilibrium. This is referred to as markets with
diminishing returns in standard economic theory. These markets are in
fine-tuned dynamic equilibrium, in which no change takes place in the
composition of the market. The law of diminishing returns ensures that
this equilibrium state always will be reached and that this state always
is the best choice.

Definition 11.1 Law of Diminishing Returns

If additional units of one production factor are employed, with all
other held constant, the output generated by each additional unit
will eventually decrease (Bannock et al., 1998).

The traditional view in microeconomics is that the market is governed

by the balance between supply and demand and that all competitors
have perfect knowledge about the market. In this simple market theory,
an evolving market (e.g., the mobile phone market) will end up in a
single predetermined equilibrium state, regardless of initial conditions
and events taking place as the market evolves. In accordance with the
law of diminishing return, any deviation away from this equilibrium
will quickly be counteracted by negative feedback such that the market
returns to the equilibrium state. If a new competitor is added to the
market, the market will initially be out of balance but will soon reach a
new equilibrium state uniquely determined by the new number of
competitors.
In 1990, Brian Arthur wrote an article in Scientific American in
which he claimed that many dynamic markets will not settle in an
equilibrium state predicted by conventional economic theory (Arthur,
1990). This is the case for products where there is positive feedback
from the market (or network effects; see ► Chap. 8) resulting in
increasing returns (sales stimulate more sales). As explained above,
diminishing returns imply that the market contains a single stable
equilibrium state. The evolution of markets with increasing returns due
to market feedback is different. Already in 1986, Arthur and coworkers
had shown mathematically—using a method called Polya urns—that, in
general, several equilibrium states may exist in systems with positive
feedback (Brian et al., 1986). This implies that these systems may settle
in an arbitrary equilibrium state depending on initial conditions or
internal and external forces acting on the system. This is called path
dependence or “history matters” (Liebowitz & Margolis, 1995). The
chosen equilibrium is quite arbitrary and may not be the best choice.
This is contrary to standard economic theory postulating that the
market’s choice is always best.

Definition 11.2 Path Dependence

Path dependence means that the path of evolution of the market of a
good depends on the initial state of the market (e.g., number of early
adopters), network effects (e.g., bandwagon effects), and external
events taking place during the evolution of the market (e.g., product
visibility and searching costs). The different paths may end up in
different equilibrium states.

Network effects are common in the digital economy. Therefore, it is

reasonable to assume that path dependence will be common in digital
markets and that these markets may end up in one out of several
equilibrium states. There is no universal rule by which the market picks
any such state as in standard economic theory: the choice is quite
arbitrary. The path the evolution will follow depends on customer
preferences, random external events, actions taken by stakeholders,
and the timing of these random events.

Definition 11.3 Law of Increasing Returns in Classical

Economic Theory
The law of increasing returns in classical economy states that the
returns from one period to the next are more than proportionate.
This is also referred to as economies of scale (Bannock et al., 1998).

The law of increasing returns implies that if the returns during a period
T are R, then the returns during the next period T are larger than R. One
common cause of increasing returns in the production industry is that
the revenues per unit produced increase because the production cost
per unit is reduced as the production volume increases (at least up to a
certain point). Sometimes this law is referred to as the law of
diminishing cost.
The law of increasing returns must be treated differently in the
digital economy since the cost of production is of no relevance for most
digital goods. As explained in ► Chap. 6, the marginal cost of digital
goods is zero so that the cost of producing one unit of a digital good is
also zero. In the digital economy, increasing returns, therefore, usually
mean that the number of users adopting a digital good during a period
T is larger than the number of users who adopted the good during the
previous period of length T. Returns do not refer to revenues in terms
of money or valuables since, because the marginal cost is zero, the
revenue per user may also be zero (ARPU = 0; see ► Chap. 6). As
explained in ► Chap. 9, increasing returns are driven by positive
feedback from the market, or in other words, increased market
advantages generate further advantages.
A different definition of increasing returns is then appropriate for
the digital economy.

Definition 11.4 Law of Increasing Returns in the Digital

Economy
Increasing returns in the digital economy is generated by positive
feedback; that is, if a company or product gains some advantage (e.g.,
increasing number of customers or more sales), it gains further
advantages. These advantages may not generate direct revenues for
the company, though they may generate indirect revenues (Arthur,
1990).

Different forces may drive competition in different directions. Such

forces may be:
– External forces, for example, where one of the candidates is more
visible and easier to find than the competitors (e.g., VHS vs Betamax,
where more VHS than Betamax recorders were displayed in the
shops)
– Internal forces, for example, network effects (bandwagon effects,
word of mouth, “likes” buttons) in which people’s preference for a
product increases as other people buy it (e.g., Facebook vs Myspace)
The outcome is that new customers are more likely to choose one of
the suppliers in favor of the competitors and that there eventually is a
net churn of users from the competitors to this supplier. In some cases,
this leads to the situation where one of the competitors captures the
whole market. This is called a winner-takes-all market.
2 Competition and Path Dependence
Competition in digital markets with zero marginal costs (e.g., Facebook
vs Myspace) or between technologies designed to different standards
(e.g., VHS vs Betamax) may be subject to path dependence. Usually, only
one of the competitors will survive so that the number of possible
equilibrium states is equal to the number of competitors, for example,
two for the VHS vs Betamax. It is not possible to predict beforehand
which of the competitors will be the winner.
The evolution of multisided platforms is a little more complex.
Facebook is a multisided platform where one of the business sectors is
social networking services. In the competition with Myspace, arbitrary
events generated a path in which Facebook ended up as a de facto
monopoly for social networking services (see ► Case Study 11.2).
Facebook is not a de facto monopoly in the other market segments of
the platform. In the advertising market, it competes with several other
companies, most notably, Google. These market segments will end up in
a state which, for Facebook, depends on the number of users of the
social networking services and, for Google, the number of people using
the search engine. Neither of them will become a monopoly in the
advertisement sector since both are important advertisement channels
targeting different user groups.
One of the most analyzed examples of the competitive war between
technological standards is the competition between VHS and Betamax.
The case is also used as pedagogic example of path dependence
because it illustrates in a simple way how path dependence may arise
and lead to a winner-takes-all situation (see the Wikipedia article for
more details (Wikipedia, n.d.)).

Case Study 11.1 Videotape Format War

VHS
and Betamax were two competing standards for video cassette
recorders (VCR
) in the late 1970s and the early 1980s. They were
incompatible standards since cassettes designed to the VHS
standard did not work with Betamax and vice versa. After intense
competition, it became clear in the early 1980s that VHS won the
videotape format war and eventually captured 100% of the market.
 Betamax was developed by SONY and released on the consumer
market in May 1975 in Japan and on the US market the following
November. VHS was developed by Matsushita (now Panasonic) and
released in 1976 in Japan and 1977 in the USA. The market for VCRs,
offering a new form of home entertainment, grew quickly.
Betamax had the first mover advantage since it was the only VCR
available in the USA during the first year. Introduction of VHS in the
USA in 1977 triggered full competition between Betamax and VHS.
Standard microeconomic theory predicts that both standards would
prevail and share the market. However, it turned out that VHS and
Betamax were operating in a market with strong network effects
which, eventually, became a winner-takes-all market.
The VHS cassette could record longer TV shows than the
Betamax cassette due to its larger size. Of particular importance was
the fact that VHS could record a complete football match (up to 3
hours) on a single cassette. This, combined with lower prices, shifted
the market share leadership from Betamax to VHS in the late 1970s.
As a secondary effect, the number of VHS recorders on display in
retailer shops gradually increased, resulting in a bandwagon effect in
favor of VHS.
This led producers of movies and other content to favor VHS. In
the beginning, content producers made their titles available on both
Betamax and VHS; however, when VHS took the market lead, they
gradually stopped producing content for Betamax, strengthening the
position of VHS. Finally, VHS was locked into the path of dominating
the VCR market entirely.

3 Impact of Churning
In markets with several competitors offering similar services, users
may, from time to time, change their affiliation with one supplier for
another. This is called churning. Churning is common in the mobile
phone market. Churning causes fluctuations in the market shares of
each mobile network operator but has kept the distribution of average
market shares rather stable over long periods of time. This is so
because the churn from one operator to another has on average been
equal to the churn in the opposite direction.
 In other markets—for example, VHS vs Betamax and Facebook vs
Myspace—this is not the case, and a net churn in favor of one supplier
takes place.
► Box 11.1 contains a simple mathematical model for competition
between two technologies (e.g., VHS vs Betamax) showing the impact
churning has on the growth and decline of the two technologies.

Box 11.1 Mathematical Model for the Temporal Evolution of

Markets with Churning
This is a simple mathematical model for the temporal evolution of a
winner-takes-all market (e.g., VHS vs Betamax). For simplicity,
assume that the adoption rate (p) for new customers is the same for
both technologies. This does not change the validity of the
arguments: it just makes the computation simpler.
There is a bandwagon effect causing a net churning flow from
technology 2 (e.g., Betamax) to technology 1 (e.g., VHS). The model is
shown in ◘ Fig. 11.1.

Fig. 11.1 Mathematical model with churning from technology 2 to technology 1. (Authors’
own figure)

The differential equations for the dynamics of this system are:

in which dA/dt and dB/dt are the change in the number of users of
technology 1 and technology 2, respectively, N − A − B is the number
of users that have not adopted any of the technologies at time t, and
rB is the flow of churners technology 1 receive from technology 2.
Since p is the adoption rate, p(N − A − B) is the total number of new
adopters per unit time adopting either technology 1 or technology 2.
Adding the two differential equations gives:

This is a separable differential equation in A + B with solution:

Inserting this in the second equation yields a linear differential

equation for B:

with solution:

This gives for A:

◘ Figure 11.2 shows the evolution of the market for p = 0.5 and r
= 0.1. The ordinate is the market share (A/N, B/N), and the abscissa
is the number of years after the technology was introduced.
Fig. 11.2 Market evolution of two competing technologies with churning. (Authors’ own
figure)

In this model, there is a net churning from technology 2 to

technology 1, and technology 1 captures the whole market in the end.
If the net churning had been from technology 1 to technology 2, then
technology 2 would have captured the whole market. The alternative
the market chooses cannot be predicted by standard economic
theory. Both alternatives are stable equilibria, and which of them is
finally chosen is path dependent.
The theoretical market evolution shown in ◘ Fig. 11.2 is not too
different from the actual evolution of the VHS and Betamax markets
as described in ► Case Study 11.1. After about 4 years VHS had
captured almost 60% of the US market, and after 13 years VHS
controlled more than 90% of that market.
4 Path Dependence and Lock-In
Path dependence may cause strong vendor lock-in in which one
supplier dominates the market entirely. In such markets, it is almost
impossible for new vendors to enter the market to offer a competing
product. VHS vs Betamax is such a case. Another case is Facebook vs
Myspace.
The process of lock-in in a market shared by two services is
illustrated in ◘ Fig. 11.3. Initially, Service A has a small lead in market
size compared to Service B. At some point in time, an external event
boosts the popularity of Service B. This event may be caused by a
successful advertisement campaign for Service B, word of mouth in
which people start sharing positive experiences on Service B, or new
features added to Service B making it more attractive. It is hard to
predict what triggers a positive feedback event. However, the event
gives Service B a boost in market size, and as network effects trigger
the users of Service A to churn to service B, the market size of service B
increases further, while that of service A becomes smaller. Eventually,
Service B will capture the whole market. This is called lock-in if it,
afterward, is almost impossible for anyone to compete with service B
(see ► Chap. 12).
Fig. 11.3 Path-dependent evolution with one positive feedback event. (Authors’ own figure)

As an illustration of this concept, the competition between

Facebook and Myspace is described in ► Case Study 11.2.
A succession of several events may alter the market, as shown in ◘
Fig. 11.4. In this example, a positive feedback event boosts the
popularity of Service A (positive feedback event 1). The increased
number of customers overloads the capacity of Service A, resulting in a
negative network effect eventually reducing the number of users of
Service A (positive feedback event 2). Service B may first lose users to
service A, but later receive users from Service A when the popularity of
Service A declines. The popularity of Service B is also boosted by
positive feedback events, for example, after having introduced new
features.
Fig. 11.4 Path dependent evolution with two positive feedback events. (Authors’ own figure)
 Case Study 11.2 Facebook vs Myspace
Facebook vs Myspace is a recent example of path dependence. Either
of the two companies could have ended up as the market leader.
However, early events triggered Facebook to take the lead and
suppress the future growth of Myspace. Myspace was inaugurated on
August 1, 2003, while Facebook was used for the first time on
February 4, 2004. The early versions of Myspace and Facebook were
rather similar from a user’s point of view. Myspace had a market lead
on Facebook until 2008, most likely caused by being first to the
market. There is no evident reason why Facebook should have taken
the lead in 2008. One possible explanation is that Myspace was a
less-flexible service, building most of the Myspace content in-house,
and focusing on music and entertainment. On the other hand,
Facebook had a more open and flexible platform that allowed third-
party providers to create content. They also focused on a social
networking experience in which the users themselves created
content for one another.

5 Conclusions
We have seen that several business cases in the digital economy cannot
be explained by standard microeconomic theory. Microeconomic theory
postulates that the market always ends up in a single predetermined
equilibrium. Two cases analyzed in this chapter, VHS vs Betamax and
Facebook vs Myspace, both end up being a winner-takes-all market.
Either of the competitors could have won the market war so that there
are two stable market equilibria in this case. That Facebook and VHS
won the competition is just luck and help from external events working
in their favor. The evolution of these markets is said to be path
dependent.
Path dependence is caused by positive feedback from the market
stimulating growth of some competitors resulting in loss of market
shares for other competitors. Sometimes, one of the competitors will
capture he whole market. As we have seen in this and previous
chapters, positive feedback is common in the digital economy.
Therefore, it is particularly important to be aware of path dependence
and the influence it may have on both establishment of new businesses
and competing with other businesses in the digital economy.

Questions
1.
AltaVista was launched in 1996 and Google was introduced on  
the international market in 1999. What happened to the two
search engines? (See, e.g., Wikipedia articles on search engines).
2.
Can the “likes” button trigger path dependency? Explain.  
3.
May path dependence explain why similar businesses, e.g., car  
dealers, tend to agglomerate geographically?

Answers
1.
AltaVista was inaugurated in December 1995, while Google  
appeared on the international market in 1999. AltaVista had an
obvious first mover advantage over Google. Already in early
1996, 45% of the users preferred AltaVista for web search. In
2000, about 1 year after Google started operation, AltaVista had
still 17% of the market, while Google had only 7%. Since then,
the usage of AltaVista declined rapidly and soon more or less
disappeared from the market, while Google increased their grip
on the market and soon became the leading search engine,
currently, with a market share of more than 60%. The reasons
were that Google offered a more efficient search algorithm
(causing positive network effects changing the path of evolution
in favor of Google) and that AltaVista was taken over by Yahoo!
(external path-changing event where the brand name
disappeared from the market).
2. Yes. The “likes” button is a positive feedback mechanism from
the market where a positive response may increase the  
popularity of a Facebook page (positive network effect), while a
negative response may reduce the popularity of the page
(negative network effect). See ► Chap. 9 for definitions of
 positive and negative network effects and positive and negative
feedback.
3.
Several factors causing agglomeration have been suggested:  
– If there are several similar vendors selling similar products
(e.g., cars, shoes), more customers are attracted to the area
and hence increasing the sales for each vendor.
– It may be easier to attract workers with specialized skills to
the region.
– It may stimulate cooperation reducing development costs and
enhancing research.
– It is more likely that easily available support facilities will be
established in the same area.
These factors may then generate positive network effects
attracting new businesses to the same area. It has been claimed
that Silicon Valley grew up in this way.

References
Arthur, W. B. (1990). Positive feedbacks in the economy. Scientific American, 262, 92–99.
[Crossref]

Bannock, G., Baxter, R. E., & Davis, E. (1998). The penguin dictionary of economics (6th ed.).
Penguin Books.

Brian, A. W., Ermoliev, Y. M., & Kaniovski, Y. M. (1986). Strong laws for a class of path-dependent
stochastic processes with applications. In V. I. Arkin, A. Shiraev, & R. Wets (Eds.), Stochastic
optimization (Lecture notes in control and information sciences) (Vol. 81). Springer.

Liebowitz, S. J., & Margolis, S. E. (1995). Path dependence, lock-in, and history. Journal of Law,
Economics and Organization, 11(1), 205–226.

Wikipedia. (n.d.). Videotape format war.

Further Reading
Arthur, W. B. (1994). Increasing returns and path dependence in the economy. University of
Michigan Press.
[Crossref]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_12

12. Lock-In and Switching Costs

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Vendor lock-in – Customer lock-in – Proprietary lock-in –

Switching cost

Learning Objectives
After completing this chapter, you should be able to:
– Identify and classify lock-in mechanisms associated with a
digital technology or service.
– Explain how network effects may cause lock-in and sometimes
also winner-takes-all situations.
– Identify whether or not lock-in effects can be eliminated by
market regulations.

12.1 Introduction
► Chapter 9 explains why network effects may cause lock-in in certain
markets, and ► Chap. 11 illustrates how path dependence generated by
network effects may drive the market into such states. This chapter
considers more closely mechanism that can be deployed to acquire and
maintain lock-in.
 Definition 12.1 Lock-In
Lock-in incorporates all mechanisms that a company may use to keep
its customers by establishing barriers to prevent customers to
switch to another supplier. This is referred to as vendor lock-in
(focusing on the vendor instigating lock-in), customer lock-in
(focusing on the customer being locked in), or proprietary lock-in
(focusing on the product or service into which lock-in takes place).

We will use the term lock-in for all three terms since they only display
different perspectives of the same phenomenon.
Acquiring new customers is expensive since it often involves
intensive marketing and expensive price campaigns. Therefore, the
vendor must do whatever is possible to keep the customers. Churning is
defined to be the act that a customer abandons the service offered by
one provider for a competing service offered by another provider.
Vendors want to reduce churning as much as possible by exploiting
various lock-in mechanisms, for example, by making it expensive or
inconvenient for users to switch to other suppliers.
Note that lock-in may lead to de facto monopolies such as in the
standards war between Betamax and VHS (see ► Chap. 11). Lock-in
combined with path dependence is also the reason why Facebook, and
not Myspace, became the leading social networking medium on
Internet. The wisdom is that the level of lock-in for a digital service
significantly influences the evolution of market shares, competition,
and formation of monopolies.

12.2 Switching Costs

Definition 12.2 Switching Costs
Switching costs are the direct and indirect costs for suppliers to
capture customers from a competitor (vendor switching costs) and
for a customer to switch to a new supplier of a good or service
(consumer switching costs)
. The total switching cost is the sum of the
cost for the supplier and the cost for the customer.
If the vendor’s switching cost is V and the customer’s switching cost is
C, then the total switching cost is S = V + C.
For the vendor, there are two strategies (Amarsy, 2015):
1.
To capture new customers, the vendor must make the switching  
cost for potentially new consumers as small as possible. In practice,
this means that the vendor must compensate for expenses or
inconveniences the customer may have for swapping supplier. This
makes the switching costs for the vendor high.
2.
On the other hand, the vendor must make the switching cost as  
high as possible for its own customer to discourage them from
switching to a competitor. The switching costs should be so high
that competitors are discouraged from trying to capture the
customer.
In oligopolies, this may sometimes cause conflicting strategies,
particularly, if the switching costs are purely monetary as illustrated in
► Example 12.1.

Example 12.1 Competition in the Mobile Phone Market

Until about 2010, it was common for mobile operators to sell mobile
phones to new customers for a much lower price than the actual
market price for mobile phones to capture new customers. This
reduced the switching costs for the consumer but increased them for
the supplier. The total switching cost was unchanged. Since the
mobile market is an oligopoly, the mobile operators were forced to
play a prisoner’s dilemma game, in which all suppliers were
compelled to use the same pricing strategy (see ► Chap. 13 where
prisoner’s dilemma is explained). If not, they would capture fewer
new customers, and the cost for own customers to switch to a
competitor would be small. When the market for smartphones
approached saturation, this practice was terminated, and the
consumers had to pay the market price for smartphones. One reason
for the new strategy was, of cause, that the operators had realized
that subsidizing the phones was a bad strategy in markets that were
saturated, as the mobile market in Europe had become at that time. In
an oligopoly market where there are few new customers to capture,
 this strategy will not stimulate growth. The strategy will reduce the
expenses of the customers but increase the costs and reduce the
revenues of the mobile operators.

The switching cost for the consumer is composed of several elements

such as fees for terminating a subscription (now mostly nonexistent
because of market regulation), lost advantages (conditional savings and
discounts), additional work (installation and training of staff), possible
loss of information (incompatible formats), hidden costs (additional
equipment or functionality not included in the offer), inconvenience
(updating cooperating systems, customers, or address lists), and
surprises (the offer is not as good as promised). Some of the switching
costs are direct costs (e.g., exit fees, training, or additional equipment),
and some are psychological, emotional, or social (“the pain of losing an
advantage is stronger than the pleasure of gaining the same advantage”
(Kahneman & Tversky, 1992)).
The switching costs for the vendor may be introductory price offers
(first month free of charge), discounts on equipment (mobile phone for
a low price), training assistance (free training course for key
employees), or free additional features (antivirus protection or backup
storage).

12.3 Lock-In Mechanisms

A company may employ different strategies to lock in its users. Here are
some of the main strategies (Shapiro & Varian, 1999).

12.3.1 Spare Parts, System Updates, and Maintenance

Systems where availability and dependability are critically important
(e.g., power distribution, telecommunications, airlines, and large
computer systems), it must be easy to obtain spare parts to keep the
system up and running. The company may have a separate store of
spare parts and skilled maintenance personnel, or they may have a
contract with the equipment manufacturer to supply spare parts and
repair at short notice. All this costs money. This leads to lock-in for the
system owner because switching to a new supplier while the existing
system is still in operation means a double set of spare parts (and
related maintenance contracts).

12.3.2 Training
The situation is much the same as for spare parts. A new system—for
example, a new type of aircraft—may require extensive training of staff
for operations and maintenance. The same applies for a change of
computer platforms from, for example, Microsoft to Mac, in which lock-
in mechanisms usually favor Microsoft since they have the biggest
market share.

12.3.3 Incompatibility and Compatibility

There are several examples where incompatible formats, procedures, or
technologies may lead to lock-in. Some examples of lock-in caused by
incompatibility are in ► Example 12.2. A more complex case is
backward compatibility since it sometimes is used to avoid lock-in,
while in other cases, backward compatibility creates lock-in. ►
Example 12.3 discusses the effect backward compatibility may have on
lock-in.

Example 12.2 Lock-In by Incompatibility

Lock-in by incompatibility is best exemplified by the lock-in strategy
of Microsoft known as embrace, extend, and extinguish (EEE) (See
Wikipedia article: Embrace, extend, and extinguish):
– Embrace: base the product on a public standard also used by
competitors.
– Extend: add features not in the standard that cause interoperability
problems with other products based on the same standard.
– Extinguish: because of dominant market shares, the new product
may become a de facto standard squeezing out competitors not
having access to the extensions. The users are then locked into a
Microsoft product.
App Store and Google Play Store are incompatible marketplaces
for software applications for smartphones, and since the platforms
are incompatible, the user cannot switch from an Android telephone
to an Apple platform (or vice versa) without losing downloaded apps
—the two marketplaces require incompatible types of smartphones.
Skype is an over-the-top technology incompatible with voice-over-
IP (VoIP) technologies such as the technology used in 4G and 5G
mobile systems. Both parties in a conversation need Skype software
to encode and decode the speech samples. Skype is indistinguishable
from other web services and is therefore not subject to specific user
charges. This cause lock-in to Skype services, particularly, for long-
distance telephone calls. Skype is just one example of incompatible
file formats causing lock-in to a particular technology.

Example 12.3 Avoiding and Creating Lock-In by Backward

Compatibility
One important aspect of standardization is built-in backward
compatibility. This means that a new version of equipment or
software can operate smoothly together with earlier versions. In
many protocols (e.g., IP), the first information element in the format is
the version number, allowing the computer to switch to the
appropriate software for reading remaining parameters (e.g., IPv4 or
IPv6). Equipping computers with software for both IPv4 and IPv6,
lock-in to a particular network technology is avoided.
Backward compatibility of mobile phones is beneficial both for
network operators and users—network operators may smoothly
build out the network with new technology, and the users need not
buy new phones to continue using the network. Some backward
compatibility in mobile systems is built into the specifications (e.g.,
offering GSM, 3G, 4G, and 5G in the same network). Some backward
compatibility is implemented by the manufacturers so that the same
phone can be used in networks using different network technologies
(e.g., 3G and the American CDMA standard and 2G, 3G, 4G, and 5G
networks where the smartphone can attach to either of these
technologies), hence making roaming independent of network
technology. Other examples of backward compatibility are game
consoles, reuse of fixed network technologies in mobile networks, and
radios (DAB
radios can receive both digital signals and analog FM
signals). In these examples, backward compatibility does not cause
 lock-in; on the contrary, backward compatibility reduces or even
eliminates switching costs.
Backward compatibility is also related to the “grandfather clause”
where an existing standard (or system) continues to be in force after a
new standard (or system) replacing it has been put into operation.
The “grandfather” of mobile communications, GSM, is still used
together with 3G, 4G, and 5G mobile systems. The MARISAT system,
owned by COMSAT General, for satellite communications to ships was
adopted by the intergovernmental organization INMARSAT as its first
satellite system in parallel with systems developed by the
organization itself to allow smooth transition from one technology to
another and to avoid unnecessary lock-in to the old technology.
Microsoft products (e.g., Windows, Word, PowerPoint, and Excel)
support backward compatibility. The backward compatibility of
Microsoft products is beneficial for users of computer systems
because it guarantees that documents written several years ago can
still be read and modified using the current standard. The motive is
obviously to make the Microsoft products more attractive and to
increase the lock-in of Microsoft users.
Backward compatibility may then in some cases cause lock-in
(e.g., to Microsoft products), while in other cases, backward
compatibility reduces the switching costs of the users and does not
lead to lock-in (e.g., mobile communications networks).

12.3.4 Potential Loss of Information

If a company changes its computer platform, the new system may not
support the old software. This may result in loss of information or
instigate major work to be done to convert the software. Even if this is
not the case, the belief that information may be lost for some
unspecified reason may be enough to cause lock-in. Database systems
tend to grow because of the fear that the removal of an old database
may cause loss of information or that deleted information may turn out
to be useful after it has been removed.

12.3.5 Investments and Economic Lifetime

Most expensive equipment and systems usually have long technical and
economic lifetimes and cannot be replaced without substantial cost.
The manufacturer of the system may then earn money on maintenance,
by upgrading the system with new functionality and by expanding the
system. This is the case, for example, for telecommunications
equipment—long equipment lifetime (GSM is, for example, almost 30
years old and still in operation), together with expensive training and
maintenance, may be an efficient lock-in of network operators over
long periods of time.

12.3.6 Difficult-to-Terminate Contracts

Suppliers of services may impose real or perceived penalties upon
customers trying to terminate the contract. Examples are:
– Loss of interests on savings.
– Paying back expenses for expensive training if the employer decides
to quit the company before the end of a contractual period.
– Binding time for mobile subscriptions; see ► Example 12.4.

Example 12.4 SIM-Lock

SIM-lock implies that the smartphone will not accept a SIM from a
different mobile operator either permanently or for a limited period.
This is referred to as SIM-lock. Switching to another operator then
implies that the customer must buy a new phone or pay a fee to the
original operator to unlock the SIM. In the early years of mobile
communications, such subscription contracts were common. This was
one reason why operators heavily subsidize mobile phones, thereby
reducing the direct switching costs for subscribers churning from a
competing operator. This practice is now usually regulated by the
authorities stimulating competition and reducing the power of
dominating operators. In some countries, binding is not allowed at all,
while in other countries, binding is allowed for a limited time (e.g., 1
year). A supplement to SIM-lock is to offer, as part of the subscription
contract, non-transferrable insurance of the phone in case of damage.

12.3.7 Loyalty Programs

Loyalty programs are used by, for example, airlines, hotels, and retailers
to stimulate customers to purchase their services rather than those of
competitors. Club membership is similar. Club members may get access
to goods and services not available for non-members, or they may gain
other advantages (e.g., lower price, priority, or gift certificates).

12.3.8 Bundling
Several suppliers bundle different products to make their offer more
attractive, for example, offering television, on-demand streaming
services, Internet access, telephony, and proprietary content (e.g., free
music lists and movies) as one subscription. The interface equipment
may also be owned by the provider or designed to proprietary
standards so that switching also compels the customer to buy new
hardware. The customer may also lose access to content.

12.3.9 Search Algorithms

This has to do with the inconvenience (cost and time) it takes a
customer to find a particular product among millions of other products.
By implementing efficient search algorithms, a company may ease the
search for products causing lock-in of the customers. Amazon and eBay
are businesses in the digital economy benefiting from efficient search
algorithms. The same applies to Airbnb, Uber, booking services, and
streaming of music, films, and video. The search algorithm is
particularly useful to identify rare products in the long tail and
stimulate sales of these products (e.g., Amazon).
Retailers may also benefit from efficient search algorithms picking
out potential customers of a product. The retailer may then target
advertisements and sales promotions directly at the customer.
Facebook, Google, and influencers lock in advertisers and marketers by
offering such algorithms. This is the reason why Facebook is so big in
the advertisement business.

12.3.10 Product Tying

Product tying implies that the customer must purchase another
product (called the tied product) in addition to the product the
customer wants to buy. In some cases, this may lead to lock-in. For
example, Apple initially sold iPhones in the USA together with a 2-year
contract with the carrier Cingular (now AT&T) locking the customers
involuntarily to this carrier (Honan, 2007). Microsoft Windows and
Internet Explorer are also tied, making the market for independent
browsers difficult because the strong lock-in of Microsoft Windows
induces strong lock-in also of Internet Explorer. Product tying is
generally forbidden in the USA, Europe, and elsewhere but is
sometimes difficult to stop (Microsoft).
Google’s suite of apps is also a tied good but, since it is free for
manufacturers of mobile phones to install the apps and for users to
download them, is not regarded by the US Supreme Court to be illegal.
On the contrary, the Court found that the practice of Google was
beneficial for innovation and sales of mobile devices in general (Sidak,
2015).

12.4 Lock-In Cycle

Shapiro and Varian illustrate the dynamics of lock-in using the simple
diagram shown in ◘ Fig. 12.1 (Shapiro & Varian, 1999). The lock-in cycle
is as follows:
– The cycle starts when the user chooses a new product (brand
selection phase). This can be anything from becoming a user of a
social networking service, subscribing to music streaming services,
building a new 5G mobile network, or purchasing a new smartphone.
For the supplier, it is important to be more visible in the market and
have a better reputation than the competitors. This then increases
the likelihood that this supplier is among the brands the customer
may investigate further.
– In the sampling phase, the customer may start using the new brand if
it is available without binding conditions or, in case of large
purchases, start negotiating price, amount, delivery conditions, and
other contractual conditions with one or more suppliers. The
supplier may offer inducements that stimulate the user to try the
product, for example, offer the service for free for some time or offer
free samples of the product. This may also include compensating
costs associated with switching from another supplier. On the other
hand, the customer should consider both the likelihood and
consequence of being locked into a particular product and estimate
the costs of later switching to another supplier.
– The entrenchment phase is entered when the customer chooses a
particular brand. The strategy for the supplier is to build in
 mechanisms or incentives that gradually increase the switching cost
of the customer. Note that a competitor can always compensate for
pecuniary switching costs as explained above. Therefore, the most
efficient lock-in mechanisms are those associated with irrecoverable
loss (real or imagined) of assets such as information.
– In the lock-in phase, the customer switching cost has become
substantial, discouraging the customer to switch to another supplier.

Fig. 12.1 The lock-in cycle. (Authors’ own figure)

The suppliers may use the lock-in cycle as a tool in strategic analysis
and planning. Decisions taken at each stage of the cycle may influence
the growth of the future customer base and revenues. In the brand
selection and sampling stages, it is important to attract lucrative
customers and avoid customers that generate insignificant revenues. In
the entrenchment phase, the supplier must build in mechanisms that
discourage the customer to switch to another supplier, for example,
making the customer dependent of information and procedures that
cannot be provided by the competitors.
The diagram may then be used to visualize how path dependence
occurs and traps the customer in different lock-in states.
12.5 Network Effects and Lock-In Cycle
Network effects, as explained in ► Chaps. 9 and 11, are among the
strongest lock-in mechanisms, often resulting in de facto monopolies.
This is usually not a vendor-controlled lock-in, but one that is caused by
natural, strong positive feedback from the market in favor of a
particular service or technology (e.g., Facebook vs Myspace).
The most prominent example of a company benefitting from strong
network effects is Facebook. Facebook has grown into a monopoly
because of strong network effects associated with the proficiency of
effective formation of groups of “friends” and rapid dissemination of
information within the groups. It has become virtually impossible for
other suppliers to launch a similar or better service. To do so, the
supplier must offer something that gives the users better experiences
and must be able to build up communities at least as professionally as
Facebook. It must also be possible for the users to move at least part of
their Facebook content to the new website to switch to the new service
provider. Otherwise, the users may lose information they have built up
over time. For the users, loss of information may be a strong reason for
not switching to a similar service offered by another supplier.
YouTube attracts users—both viewers and publishers—because of
the popularity of the service (bandwagon effect) and strong network
effects associated with reviews, recommendations by other viewers,
and ratings. This also leads to lock-in since it is both difficult and
expensive for competitors to build up a competing service that will give
users access to such a volume of video content and provide so much
visibility to publishers of new video material.
Lock-in to a technology is also often caused by network effects. The
two most quoted examples are the adoption of the QWERTY keyboard
(see ► Box 12.1) and VHS as a video cassette standard (see ► Chap.
11).

Box 12.1 The QWERTY Keyboard

Most of the keyboards in use today follow the QWERTY layout. This
layout was generally adopted in the late nineteenth century after the
success of the Remington 2 typewriter. The specific QWERTY layout
was selected to make the typewriters work as smoothly as possible
and to avoid jamming of the metal bars in the machine when two
letters were typed in fast succession. Enforcing a standard, such as
the QWERTY layout for keyboards, is beneficial—though not
necessary—for both keyboard users and keyboard manufacturers.
However, since the QWERTY layout became dominant, it was hard for
competing layouts to enter the market since switching costs built up
as more and more users adopted and were trained for the QWERTY
layout. The layout is shown in ◘ Fig. 12.2.

Fig. 12.2 The original QWERTY layout (By C.L. Sholes - U.S. Patent No. 207,559, Public
Domain). Note that the “0” and “1” is intentionally missing to simplify design. The “0” can be
reproduced as a “O,” while a “1” can be reproduced as an “l” or an “I.” (Source: Public
Domain, ► https://​en.​wikipedia.​org/​wiki/​QWERTY#/​media/​File:​QWERTY_​1878.​png)

The competing DVORAK layout (◘ Fig. 12.3) was patented in 1936

by Dr. August Dvorak. It is, by many, believed to be superior to the
QWERTY layout, in terms of typing speed. However, when it was
launched, it failed to get any market foothold due to high switching
costs and the lock-in of QWERTY. Since the QWERTY layout had a
dominating position in the market, it was hard to convince both users
to learn and manufacturers to produce keyboards with the DVORAK
layout. Today, almost all PCs, laptops, and smartphones use the
QWERTY layout.
 Fig. 12.3 The original DVORAK keyboard layout. (Source: Public domain, ► https://​en.​
wikipedia.​org/​wiki/​Dvorak_​keyboard_​layout#/​media/​File:​KB_​United_​States_​Dvorak.​svg)

The choice of the QWERTY keyboard layout as a standard is also

an example of path dependence. The early decision to adopt the
QWERTY layout, which was perfectly logical at that time, locked
keyboard designs into a path which would be extremely hard, if not
impossible, to leave. Adopting the DVORAK keyboard layout would
have been better for the society at large (though this conjecture is
contested by, e.g., S. J. Liebowitz and S. E. Margoliz (Liebowitz &
Margoliz, 1994)); however, the switching costs are currently too big
for revitalizing DVORAK.

12.6 Lock-In and Market Regulations

One of the most important aspects of market regulation is to reduce the
lock-in capabilities of service providers to enhance competition.
Examples of such regulations include:
– Forbidding binding time for subscriptions (e.g., in
telecommunications)
– Prohibiting loss of advantages (e.g., in insurance)
– Obligating number portability in telecommunications
The last item may require some explanation. Both fixed and mobile
telephone numbers have historically uniquely identified both the
operator and the subscriber connected to the network. If there is more
than one telephone operator in a country, each operator is, therefore,
assigned different series of national telephone numbers. In this regime,
the subscriber will get a new telephone number if he or she moves the
subscription from one operator to another. This is inconvenient for the
subscriber and causes lock-in since it may entail that hundreds of
friends and other contacts must be updated about the new number. To
avoid this type of lock-in, the regulatory authorities in several countries
introduced number portability as a mandatory requirement around
2000. This implies that the subscriber keeps the same number if the
subscription is moved to another operator. The networks must then be
updated with facilities that route the calls to new destinations without
any involvement of the users of the telephone network.
In telecommunications, some providers, particularly, incumbent or
dominating operators, are facing a particular form of lock-in, namely,
that they must continue to offer network services long after these
services have ceased to be profitable. This is usually a public duty
enforced by the authorities. Network providers must, for example, still
offer fix telephone services even though more and more of the users are
switching to mobile phones as their only telephone subscription, and
mobile network operators must still support GSM even though they
offer full national coverage with 4G technology.
◘ Table 12.1 summarizes how regulation can be applied to avoid
lock-in and undesirable market behavior. The effect of the switching
barriers is also indicated in terms of two broad categories:
– Economical barrier implies that there are considerable economic
expenses associated with crossing the barrier.
– Psychological barrier implies that the major switching cost is
associated with lack of knowledge of the outcome of the switching,
for example, fear for losing information built up over time.
Table 12.1 Lock-in and regulations in the digital economy. (Authors’ compilation)

Lock-in Regulation Switching barriers if Regulatory actions

mechanism not regulated
Spare parts Impossible Expensive for the  
and customer
maintenance
Training Impossible Expensive for the  
customer
Lock-in Regulation Switching barriers if Regulatory actions
mechanism not regulated
Incompatibility Some Expensive for the Interoperability is required by law in
regulations customer unless in certain cases (telecommunications
but mostly regulated cases networks) and between ASPs and
unregulated ISPs
Loss of Impossible May be expensive for  
information customer; in many
cases psychological
barriers
Economical Impossible Expensive for  
lifetime customer
Contracts Possible May be expensive for May be regulated by law (SIM-lock
customer and number portability)
Loyalty Possible Psychological barrier Some loyalty programs may be illegal
programs by law
Bundling Possible Psychological barrier May be illegal by law in some cases,
e.g., bundling of ASP services
provided by market dominating ISPs
Search Possible Convenient for the Not regulated
algorithms customer
Product tying Possible May be impossible to Generally forbidden but there are
escape in many cases exceptions

In some cases, it is impossible to regulate the market such that the

switching barriers disappear, for example, expenses associated with
spare parts, training, possible loss of information, and equipment
lifetime. In other cases, switching costs may be reduced or eliminated
by appropriate regulations.
– Termination of contracts without any expenses or inconveniences for
the customer is in many cases made compulsory by regulations, for
example, forbidding or limiting the use of SIM lock or requiring that
advantages gained as customers are transferrable to new providers
(e.g., in the insurance business).
– Bundling of services is also regulated in several countries. One case is
associated with incumbents, that is, operators that were state
monopolies before the general liberalization of telecommunications
in 1998. To avoid that the incumbents misuse their market power,
restrictions may be put on how these companies bundle their
 services in subscription packages, for example, combining telephony,
Internet, mobile phone, and television in one subscription and
thereby capturing all the individual markets. The physical
connection, called the” local loop,” from the subscriber to the local
exchange is usually owned by the incumbent. To allow other
operators to access the subscribers without digging new cable ducts,
the market regulations require that the incumbent allowed other
operators to use the incumbent’s connection. This is called local-loop
unbundling (LLU). LLU became mandatory in the USA in 1996
(Telecommunications Act of 1996) and the European Union in 2001.
– It is difficult to regulate markets with strong network effects. The
reason is that the network effects tend to create de facto monopolies.
If antitrust legislation is used to divide such monopolies into two or
more independent competitors, the same network effects are again
likely to turn this market into a new monopoly with only one of these
companies surviving the competition. For the same reasons, it is also
fruitless to economically or regulatory stimulate a new competitor to
enter the market. In most cases, the market dominance of the
existing monopoly is so strong that the newcomer will never grow to
become a real competitor, or if it starts growing because the network
effects are in favor of the newcomer, it may capture the whole market
and become a new monopoly squeezing the old monopoly out of the
market.

12.7 Conclusions
The key message is that lock-in is common in the digital economy. For
vendors, there are two strategies:
1.
To keep customers by making the switching costs for the customers  
(barriers to leave) as high as possible
2.
To capture customers from competitors by making the switching  
cost for customers from competitors (barriers to enter) as small as
possible
These strategies may be conflicting, particularly, in oligopoly
markets. In some cases, this may lead to a prisoner’s dilemma where
the switching barriers for the customers are low, while the vendors
carry all switching costs.
If there is a choice between two or more incompatible technological
standards offering the same services to the consumers, such as VHS vs
Betamax and QWERTY vs DVORAK, the market eventually chooses just
one of them. This is a strong case of lock-in resulting in a winner-takes-
all situation. To end up in this state depends on luck, random events,
and network effects, or in other words, the selection of final state is
path dependent.
Providers of social networking services may also end up as de facto
monopolies (e.g., Facebook). Switching barriers in favor of the provider
are built up by the users themselves by creating and joining
communities of users: the more communities a user is member of, the
larger is the psychological barrier to leave. Facebook is a good example
of a case where there are no monetary barriers against switching but
huge psychological ones.

Questions
1.
Which lock-in mechanisms tie a mobile network operator to the  
manufacturer of mobile networks?
2.
Which lock-in mechanisms tie users to mobile network  
operators?
3.
Are there any lock-in mechanisms binding users to Spotify?  
4.
Are there any lock-in mechanisms binding advertisers to  
Facebook?

Answers
1. The most important mechanisms are:
– Spare parts, maintenance, and system updates
 
– Training
– Economic lifetime
 2.
This depends on the country.  
– SIM-lock may be legal in some countries but illegal in others,
– Number portability may be mandatory in some countries but
not in others.
– Service bundling may be legal or illegal.
– The mobile operator may offer loyalty programs of some kind.
– The operator may offer cloud services (computation or
storage) that are net transferrable.

3.
If the user switches to another music provider, the user loses the  
playing lists since they are not transferrable. The user also loses
access to Spotify Codes and the opportunity to share playlists
and other content.
4.
The advantages of Facebook are the large number of users and  
the accurate knowledge about the users that the advertisers may
use to target the advertisements. This is not a strong lock-in
since there are no switching costs. The advertisers have several
other advertisement channels. If an advertiser abandons
Facebook for some reason, this may not cause significant
reduction in sales because the same customers may be reached
via other channels.

References
Amarsy, N. Switching costs: 6 ways to lock customers into your ecosystem. Strategyzer, July 27,
2015.

Honan, M. Apple unveiling iPhone. PC World, January 9, 2007.

Kahneman, D., & Tversky, A. (1992). Advances in prospect theory: Cumulative representation of
uncertainty. Journal of Risk and Uncertainty, 5(4), 297–323.
[Crossref]

Liebowitz, S. J., & Margoliz, S. E. (1994). Network externality: An unknown tragedy. Journal of
Economic Perspectives, 8(2), 133–150.
Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to the network economy.
Harvard Business School Press.

Sidak, J. G. (2015). Do free mobile apps harm consumers? San Diego Law Review, 52, 619–694.

Further Reading
Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to network economy.
Harvard Business School Press.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_13

13. Digital Monopolies and Oligopolies

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Natural monopoly – Oligopoly – Monopolistic competition –

Perfect competition – Monopsony – Oligopsony

Learning Objectives
After completing this chapter, you should be able to:
– Identify which types of markets a digital enterprise is
simultaneously serving.
– Explain how digital monopolies are formed.
– Perform strategic planning of a digital service based on
market type.

13.1 Definition of Market Types

Before examining the formation of monopolies and oligopolies in the
digital economy, we start with a short review of some basic market
types. The market type refers to the sales of a particular good or service
and not the company selling it. The company may be present in several
market types offering different products in each type of market. In
multisided markets (see ► Chap. 10), interactions exist between the
different markets served by the platform, making the business model
complex.

Definition 13.1 Monopoly

In the monopoly market, there is only one seller of the good, and
there are no substitutes that the buyers may choose instead. In
standard microeconomic theory, the monopoly may then alone
determine prices and maximize its revenues.

What makes this definition problematic in the digital economy is that

the price of several of the products offered in digital monopoly markets
is zero; that is, there are no revenues generated by the customer in
these markets. Some of the most lucrative companies in the digital
economy are monopolies in one or more of their market segments
offering services free of charge (e.g., Facebook and Google). Analyzing
these markets using supply-demand curves is meaningless since the
marginal cost of these goods are zero and, hence, independent of the
volume of production (see ► Chap. 6).
We may distinguish between three types of monopolies:
– That a company is a de jure monopoly means that the company is
protected against competition by law.
– That a company is a natural monopoly implies that the market is best
served by a single supplier rather than being shared among several
suppliers; for example, if the market is shared by several companies,
the prices of the good may be higher (and production less effective)
than if the market is served by only one company.
– That a company is a de facto monopoly means that the company has
captured almost the whole market and that the barrier to entry is so
high that new entrants are discouraged to try.
Natural monopolies and de facto monopolies are often considered
to be the same thing. However, the two market types are different. The
provider of the fixed telecommunications infrastructure may be a
natural monopoly since it may be more expensive for the users if there
are two or more competing providers offering identical infrastructure
services. It is then the cost of infrastructure that justifies the natural
monopoly in this case. Facebook is not a natural monopoly because the
existence of several competing “facebooks” will neither cause
additional costs for the customers nor reduce the versatility of the
services offered to the users. Facebook has become a de facto
monopoly; it is certainly not a natural (or necessary) monopoly.
A particular form of monopoly is the territorial cartel. These are
businesses that operate as a monopoly within a region, for example, a
country. Before the market was opened for competition, the telephone
operators were regional cartels by this definition. When the AT&T (the
US telephone operator owning most of the telephone infrastructure in
the USA) was split into regional operating companies (the “Baby Bells”)
in 1982, these companies became regional cartels. Since these
companies were not real competitors, they established a common
research center, Bellcore, to share costs concerning
telecommunications research. After liberalization of
telecommunications in 1998 in Europe (1996 in the USA), regional
cartels have almost disappeared from the telecommunications industry.

Definition 13.2 Oligopoly

In the oligopoly market, there are only a few sellers of the same
product. Actions taken by a single competitor (e.g., lowering the
price or offer complementary goods) may change the composition of
the market; that is, redistributing market shares. If there are only
two competitors, the market is called a duopoly.

One common strategy in oligopoly markets is that each firm must be

aware of actions taken by the other firms (on, e.g., price and marketing)
and respond accordingly. Actions taken by each company may then
have direct impact on prices, competitive strength, and customer
behavior. Examples of oligopolies in the digital economy are:
– Only a few mobile network operators offer services in each country
forming local oligopolies.
– Visa and MasterCard share most of the international credit cards
market.
– Microsoft and Apple are the dominating producers of operating
systems for personal computers.
– Intel and Advanced Micro Devices manufacture most of the CPUs
used in desktop computers.
– ARINC and SITA share the market for ground station-to-aircraft
communications in Europe.

Definition 13.3 Monopolistic Competition

Monopolistic competition implies that there are many sellers in the
same marketplace offering differentiated products, for example,
shoes with different designs, materials, and quality. Other examples
are restaurants and producers of cheese, soap, cars, and clothing.

The products in a monopolistic market may serve the same purpose,

but differentiation makes the products unique so that they are not exact
substitutes. The prices of products from one supplier are, in general,
independent of the prices set by other suppliers.
In the digital economy, the markets for smartphones, personal
computers, and television sets are monopolistic markets. The market
for streaming services with a few big and many small competitors may
also be regarded as a combination of monopolistic market and
oligopoly.

Definition 13.4 Perfect Competition

Perfect competition is a theoretical model that describes markets
with many sellers and buyers, where both sellers and buyers have
perfect information about prices and customer preferences. Opposed
to the oligopoly, the actions of a single competitor will not change
the composition of the market.

In a perfect market, the products are indistinguishable, there is no

transaction cost associated with the sales process, and there are no
externalities that may favor some competitors or change the market
rules. Real markets are only crude approximations of this model.
Almost perfect competition may exist in commodity markets with many
suppliers and undifferentiated goods where price is the only factor
distinguishing the products from different suppliers. In the digital
economy, perfect competition occurs in markets for digital freelance
services—webpage design, brand design, visual design, writing, and
translation services.
The above market types are related to the seller side only. There are
also markets where the market behavior is determined by the number
of buyers, that is, if there is only one or just a few buyers. There are
several examples of such markets in the digital economy. All of them are
multisided platforms (see ► Chap. 10).

Definition 13.5 Monopsony

A monopsony is a market with many sellers and only one buyer.

In the digital economy, the sellers in the monopsony are content

providers, often independent artists, who see the monopsony provider
as the only channel through which they can distribute their art. A
monopsony with both buying and selling sides controls the whole
market monopolistically, buying content cheaply and reselling it with
considerable profit.
Example of a (de facto) monopsony in a multisided market is the
video sharing service of YouTube. The three most important customer
groups are producers of video clips, viewers, and advertisers. As
monopsony, YouTube is a channel for producers to distribute video
clips to the viewers. Production and distribution of the video content is
free of charge. As a monopoly, the video clips are provided to the
viewers, also free of charge. Revenues are generated by the advertisers.
The advertising market may either be regarded as an oligopoly if only
the advertisement channels provided by the dominating stakeholders
in the digital economy are counted or as a market with monopolistic
competition if both large, small, and non-digital advertisement
channels are counted.

Definition 13.6 Oligopsony

The oligopsony is similar to the monopsony. The only difference is
that there are more than one (but just a few) buyers in the
oligopsony market, while there is only one in the monopsony
market.
The current music streaming market is an oligopsony with three major
resellers: Spotify, Apple Music, and SoundCloud. The customers in this
market are single artists and producers. Publishers of e-books are also
an oligopsony with a few publishers dominating the market. The
strategy of the oligopsony with both a buying and a selling side is more
complicated than that of the oligopoly since it competes in two
interacting markets at the same time: competing as buyer of content
and competing as resellers of the same content.
◘ Table 13.1 and ◘ Fig. 13.1 summarize the various market models
defined above. The table characterizes the markets in terms of the
number of sellers, the number of buyers, and the barriers of entry.
Table 13.1 Market types. (Authors’ compilation)

Market type Number of Number of Barriers to Examples in digital

sellers buyers entry economy
Monopoly One Many Impossible Facebook
Oligopoly Few Many High Mobile network
operators (MNO)
Monopolistic Many Many Open Smartphones
competition manufacturers
Perfect Many Many Open Digital freelance services
competition
Oligopsony Many Few High Spotify
Monopsony Many One Impossible YouTube
Fig. 13.1 Market types. (Authors’ own figure)

13.2 Formation of Monopolies

Until 1998, the telecommunications businesses in most countries were
government-granted monopolies (or de jure monopolies). Full
competition was introduced in Europe in 1998, allowing anyone to
become a network or service provider (ISP
). To prevent the incumbent
(the former monopolist) from misusing its market power built up on
historical government money, the fairness of competition was strictly
regulated by the government. The regulations impede the incumbent
from buying up competitors or forcing them out of the market with
unfair pricing or other obstructions of their business. The regulations
contain technical and commercial conditions for how newcomers can
interconnect their networks to the network of the incumbent, also
allowing the newcomers to operate as resellers or virtual network
operators (VNOs). Despite the regulations, monopolies have arisen in
the ICT businesses, especially in information service markets. These
newcomers (e.g., Facebook, YouTube, and Twitter) are referred to as
natural monopolies or de facto monopolies.
The term natural monopoly was formally defined by William
Baumol as “[a]n industry in which multi-firm production is more costly
than production by a monopoly” (Baumol et al., 1982). The natural
monopoly has 100% market share. The closely related term “de facto
monopoly” implies that the company may not have 100% market share
but will have nearly so over a substantial amount of time. Therefore, the
de facto monopoly is not a true monopoly.
Path dependence caused by strong network effects may, at one
point, work in favor of one of the competitors who eventually will
capture most of the market (see ► Chap. 11). This happened in the
competition between Facebook and Myspace, in which Facebook took
the lead in 2008. Myspace is still active, having about one million
registered users, while Facebook has 2.2 billion users and is a de facto
monopoly, as it is more than 2000 times bigger than its competitor.
Note that the same person may be a registered user of Myspace and
Facebook at the same time. Membership in one social media network
does not exclude simultaneous membership in a competing social
media network.
As explained in ► Chap. 12, lock-in implies that it is difficult for a
newcomer to capture market shares in a market dominated by a de
facto monopoly. Though Myspace is still active and offers competitive
services to Facebook, lock-in in favor of Facebook hampers Myspace to
capture significant market shares from Facebook.
Cost may be a factor in some cases. In the VCR standards war, two
incompatible standards are more inefficient as compared to one
standard. VHS and Betamax are almost identical as seen from the user’s
viewpoint provided that the same films are available on both standards
for approximately the same price. However, the filmmakers may view it
differently; they must produce two versions of the same film for two
incompatible media. This is both expensive and cumbersome for the
production side and will eventually lead to higher prices; therefore, the
market eventually develops into a de facto monopoly since bandwagon
effects, for example, that one of the products are more visible in
advertisements and, in displays in shops, may work in favor of one of
the technologies (see ► Chap. 11 for more details)
.
Sometimes, companies with large market shares—for example,
Google—may be mistaken to be de facto monopolies. In the search
engine market, Google has 92% of the market, and the rest is divided
between Bing (2.5%), Yahoo! (1.5%), Baidu (1.5%) (China), and several
other search engines with less than 1% market share each (Search
engine market share worldwide: Dec 2019–Dec 2020. StatCounter).
The web browser market is shared between Google Chrome (64%),
Safari (18%), Mozilla Firefox (4.5%), Samsung Internet (3%), and
several other browsers with small market shares (Browser market
share worldwide: Dec 2019–Dec 2020. StatCounter). In the search
engine market, Google is strictly not a de facto monopoly since there is
still considerable room for competition, and there are no strong lock-in
mechanisms that bind the user to one particular search engine. The
market shares in the search engine market have also been rather stable
for a long period of time. The browser market is certainly not a de facto
monopoly, though it is dominated by one large provider almost four
times bigger than the next largest provider.
A contributing cause of formation of de facto monopolies in the
digital economy is the increased gap between value and cost as more
and more users adopt a digital service. This is illustrated in ◘ Fig. 13.2
for digital services with strong network effects obeying Metcalf’s law
(see ► Chap. 9)
. As explained earlier, many digital services and goods
have zero marginal cost. This implies that the cost curve approaches
zero as the production volume (n) (e.g., the number of users in social
media) increases. On the other hand, the value of the digital service—
and, hence, the revenues of the company offering the service—
increases linearly as a function of the number of users (Metcalf’s law).
The result is a big gap—a value surplus—between value and cost, as
depicted in ◘ Fig. 13.2. The value gap increases as the production
volume increases and gives companies with large n a strong financial
position. This can be seen in, for example, the profit margins of many
companies producing digital services, which often range from 25% to
50%—well above the normal industry standard. Digital companies may
use this profit margin for more growth; for example, developing and
improving existing services, creating new services, and/or acquiring
competing or supplementary businesses.

Fig. 13.2 Difference in value as a function of units produced. (Authors’ own figure)

Case Study 13.1 Creation of a Monopoly: Facebook

The forerunner for Facebook, the website Facemash, was created in
2003 at Harvard University. The website was first only available for
students at Harvard College but expanded to other universities in
2004. In 2008 after almost 5 years, it became a universally available
website on the international market where users could form and
participate in social groups exchanging experiences, news,
comments, views, pictures, and other information. Already in 2009,
it had surpassed Myspace in the number of users, exceeding 500
million users in July 2010 (Zuckerberg makes it official: Facebook
hits 500 million members, 2010). The company had then become a
monopoly in the social services business: there were no real
competitors that could match it in growth and market impact.
From 2008, the growth of Facebook has been governed by strong
network effects creating lock-in with high barriers for competitors to
enter the same market. The barriers created by Facebook were not
economic but psychological. The major concern for the users is that
they may lose all information they have produced and collected, as
well as being chopped off from the network of interactions with
friends and other user they have built up.
The primary service of Facebook is free of charge for the users.
This makes the entry barriers for competitors even higher: the
competitors cannot provide the same service cheaper, leaving them
with the alternatives to pay users for joining their platform to
differentiate them on price or offer a better customer experience
with less exploitation of personal data. This will require enormous
efforts and ingenuity and is extremely expensive.
The platform is designed such that Facebook can extract
enormous amounts of data about the users such as personal data
(gender, age, geographic location, work, etc.), political preferences,
network of friends and contacts, habits, motivations, cultural
preferences, and so on and so forth. Facebook sells information
based on this knowledge to marketers and other organizations
utilizing the information for statistics, trend analysis, sociological
studies, lobbying, opinion shaping, surveillance, and other purposes.
It is this sale of personal information that generates the revenues of
Facebook.
Facebook does not meet competition on its social networking
platform: it has long ago become a de facto monopoly. The
competition Facebook encounter is on advertisements. It turns out
that it is a leading stakeholder in this respect also. As of January
2019, 94% of marketers worldwide were using Facebook for
advertising, and 74% were using its subsidiary Instagram. In
contrast, 59% were using Twitter, 58% LinkedIn, and 54% YouTube.
All other social media combined were used by less than 50% of the
marketers (Leading social media platforms used by marketers
worldwide as of January 2019. Statista).
Social media earns money on advertisements because of a strong
cross-side network effect caused by the large number of users as
explained in ► Chap. 10. This effect is particularly large for
Facebook since it has the largest number of users and, perhaps most
important, that it can discover and store more personal information
 about the users than other social media, facilitating marketers to
target their advertisements precisely at individual users.

13.3 Formation of Oligopolies

Mobile communications and streaming services are two examples of
oligopolistic markets in digital economics.
In 1992, mobile communication was deregulated in Europe, and,
stimulated by the governments, two or three operators in each country
acquired license to offer GSM services. Competition was later enhanced
by allowing resellers to buy bulk traffic from the mobile network
operators and reselling it to their own customers at a lower price than
the network operator. Another category of competitors is mobile virtual
network operators (MVNOs) owning some infrastructure such as
subscription and location management databases and gateway
exchanges but leasing radio access infrastructure from ordinary mobile
network operators. The mobile market is not big enough to support
many operators of different types so that mobile telecommunications
within a country has become an oligopoly.
There are several reasons why mobile communications is an
oligopoly market. The most important (and often overlooked) reason is
that the frequency spectrum allocated to mobile communications is
rather narrow and can only be sliced into a rather small number of slots
broad enough to support a single operator. This then limits the number
of operators that can build their own network in a region or country.
Each mobile network operator must then be granted a license for using
a particular slice of the available spectrum.
The second reason why there are so few mobile network operators
is that it is expensive to build and manage mobile network
infrastructures, in particular, since the licensing authorities may
require that the network cover a certain percentage of the population
(e.g., everyone) and not just the most profitable parts of the country.
Resellers and MVNOs require only small capital investments and are
easier to establish, and the mobile network operators are forced by
government regulations to let them buy bulk traffic or lease
infrastructure to affordable prices, thereby enhancing competition.
 Streaming services are serving two markets: the provider of
information to be streamed and the receiver of the streamed content. In
music streaming, there are a few big providers, where Spotify, Apple
Music, and SoundCloud are the most prominent. These providers are,
on the oligopsony side, trying to capture artists and record labels on
exclusive contracts and, on the oligopoly side, trying to capture
listeners using different business models. This includes offering a
combination of freemium and premium services (Spotify), creating
communities of artists and listeners (SoundCloud), and offering access
to a vast library of songs (Apple Music).
The competition between oligopolies is difficult because decisions
made by one stakeholder may have direct impact on prices,
competition, and market shares and, thus, changing the market
composition entirely. One particular problem is that the competitors
may fall into the prisoner’s dilemma trap. The prisoner’s dilemma is
one of the most studied games in game theory (See the Wikipedia
articles on “Game theory” and “Prisoner’s dilemma” for more details).
The payoff matrix for the prisoner’s dilemma game for a duopoly is
illustrated in ◘ Fig. 13.3. There are two competitors, firm A and firm B,
competing for the same customers. Both firms know that the other firm
is likely to lower the price to attract customers from its competitor. The
decision each firm is facing is then either to keep the price unchanged
or lower the price. The pros and cons for this decision are summarized
in the payoff matrix as shown in figure.
– If the two firms do not change the price, the situation remains
unchanged.
– If firm A decides to lower the price, then firm A may gain so many
customers from firm B that its revenue becomes larger than it was
before, while the revenues of firm B will drop considerably; firm B
may even be pushed out of the market.
– Firm A also knows that if firm B lowers the price and firm A does not,
firm A will face the same destiny.
– If both firms lower the price, the market size of the two firms will be
unchanged, but the revenues have become smaller for both firms
because the user pay less for the service.
Fig. 13.3 Payoff matrix for the prisoner’s dilemma game. (Authors’ own figure)

The worst outcome for firm A happens if it keeps the same price,
while firm B lowers the price. At the same time, the best outcome for
firm A happens if it lowers the price and firm B does not. So, what shall
firm A do? The likely outcome (called the Nash equilibrium) is that firm
A lowers the price and, by the same reasoning, firm B does the same.
This benefits the users, but the revenues of both firms are now lower,
and the business is less profitable.
 The discussion above applies to two competing companies but can
easily be extended to an oligopoly consisting of more than two
companies.
An iterated prisoner’s dilemma game is a game which is played
several times in succession. Price war is one outcome of iterated
prisoner’s dilemma games where each competitor tries to follow the
actions taken by the other competitors. One example is the continuous
price war between regional gasoline stations: if one station reduces the
gasoline price, then all the other stations are likely to do the same.
In the early days of mobile communications, price war forced the
operators to offer heavily subsidized mobile phones to the customers.
Since the mobile communication business is an oligopoly, all operators
had to choose this strategy; otherwise, they would soon be out of
business. The strategy had one advantage, namely, that it increased the
adoption rate of mobile phones causing the market to increase rapidly.
As the market matured, the practice changed because it simply meant
less revenue for the operators and did not create any new market
opportunities. The competition then changed, and the operators began
to offer complex subscription packets consisting of various
combinations of price, bandwidth, data volume, and other features to
differentiate one another. The market then became more like
monopolistic competition.

Case Study 13.2 The Mobile App Duopoly

Several companies offer mobile apps, but the market is dominated
by only two of them: Apple’s App Store and Google Play. The app
market is thus a duopoly.
The apps are designed for two types of smartphone technologies:
– App Store apps can only be loaded down on iPhones and some
other Apple products. The apps are running on the closed-source
iOS operating system of Apple. The market share of iPhone is
approximately 15% of the international smartphone market.
– Google Play apps can be loaded down on smartphones with the
open-source Android operating system of Google. Android phones
are produced and marketed by several independent
manufacturers. The market share of Android smartphones is
about 85%. The market share for Google Play is then 5.7 times
 larger than that of App Store since the market for apps is the same
as the market for operating system technologies.
Several of the apps are available for both iPhones and Android
smartphones. Some apps are available only for one of the
technologies. Google Play was designed as a production platform for
independent app developers. Initially, App Store was proprietary but
was soon opened for independent developers to produce their own
apps directly on the platform.
The business models of Apple and Google are different as is
evident from the statistics for mobile apps shown in ◘ Table 13.2
(Nelson, 2018). The data are for the third quarter of 2018.
Table 13.2 Mobile apps statistics. (Authors’ compilation)
  Number of mobile Mobile apps Mobile apps Revenue per
apps (millions) revenues (billion downloads download ($)
$) (billions)
App 2.2 12.0 7.6 1.6
Store
Google 2.8 6.2 19.5 0.32
Play

Both app stores contain approximately the same number of apps.

The majority of the most popular apps (e.g., Facebook, Instagram,
YouTube, Uber, and Google Maps) are available in both stores. The
table shows that Apple acquires 66% of the total revenues from the
app market (12 out of 18.6 billion $), while only 33% of all
downloaded apps are from App Store. Other statistics show that
iPhone users spend almost twice as much on paid apps than Android
users (Blair, 2019). This is one reason why Apple, despite having a
much smaller market share, earns more on apps than Google. Other
reasons are associated with different charges for developers and
share of revenues per app.
The apps duopoly is obviously much more complicated than
other duopolies since there is no simple relationship between
market shares and revenues.
13.4 Conclusions
The two dominating market forms in digital economy are de facto
monopolies and oligopolies. De facto monopolies emerge in markets
with strong positive network effects as explained in ► Chap. 9. The
monopoly builds huge barriers so that it is almost impossible for new
entrants to compete with them. Standard economic theory is obviously
not applicable to these monopolies in cases where the services are
offered free of charge for the customers since, in such cases, supply-
demand curves are meaningless.
Oligopolies are formed in markets where there is room for only few
competitors, for example, mobile network operations. In cases such as
streaming services, there may be many competitors, but just a few of
them are big enough to dominate the market. These markets may then
also be regarded as oligopoly markets, though they may become
markets with monopolistic competition if the suppliers are able to
differentiate their products. In such markets, both big and small
enterprises may live peacefully together.
Another feature that makes digital enterprises more complicated
than standard industrial enterprises is that they often are multisided
platforms serving two or more interacting markets. The digital
enterprise may then be a monopoly in one market and an oligopoly in
another. Services may be offered for free to one group of users, while
other user groups pay for the product where the price depends on the
number of users receiving free services (cross-side network effect).
Therefore, decisions taken in one market segment also depend on
decisions taken in the other market segment, rendering business
modeling and strategic planning intricate.
Monopsony and oligopsony markets have also appeared in the
digital economy, for example, enterprises offering streaming services
where the enterprise buys content from artists and resells it to their
own customers.

Questions
1. Facebook is a multisided platform offering services in several
market segments; three of them are social networking services,  
 advertisements, and third-party services. How will you
characterize each of them?
2.
The credit card market is served by two-sided platforms. Who  
are the two customer groups served by credit card companies?
What type of market do they represent?
3.
Why can we model the advertisement market of enterprises in  
the digital economy both as an oligopoly and as a market with
monopolistic competition?

Answers
1.
Social networking, de facto monopoly; advertisements,  
monopolistic competition because there are several competitors
offering different advertisement services to different user
segments (see also Question 3); third parties, oligopsony buying
content from third parties and reselling it to social networking
consumers.
2.
The customer groups are card users and merchants. Both user  
groups are buying the service from the companies. Since there
are few credit card companies, both markets are oligopoly
markets.
3.
The advertisement market in the digital economy is dominated  
by a few big companies and several small ones. The market may
then be modeled as an oligopoly since it is only the companies
with large market shares in the advertisement market that can
manipulate the evolution of the market. On the other hand, the
total advertisement market also contains a large number of non-
digital companies (e.g., newspapers and journals) reaching
different segments of the population. The advertisement market
can then also be modeled as a market with monopolistic
competition because each stakeholder may offer marketers
access to different segments of the population (e.g., different age
groups) and to special interest groups.
References
Baumol, W. J., Panzar, J. C., & Willig, R. D. (1982). Contestable markets and the theory of industry
structure. Harcourt Brace Jovanovich.

Blair, I. (2019). Mobile app download and usage statistics. BuildFire.

Nelson, R. Global app revenue topped $18 billion last quarter, up 23% year-over-year. Sensor
Tower. October 10, 2018.

Zuckerberg makes it official: Facebook hits 500 million members. TechCrunch. July 4, 2010.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_14

14. Mergers and Acquisitions

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Organic growth – Friendly takeover – Hostile takeover –

Horizontal integration – Vertical integration

Learning Objectives
After completing this chapter, you should be able to:
– Explain the difference between merger, hostile acquisition,
reverse takeover, and backflip.
– Understand the concepts of horizontal integration, vertical
backward integration, and vertical forward acquisition.
– Explain how mergers and acquisitions have shaped digital
markets.

14.1 Definitions
► Chapter 13 argued that several companies in the digital economy
progress towards de facto monopolies over time. The main reason for
this is strong positive network effects caused by positive market
feedback. Large companies have an advantage by just being large and
tend to capture even more consumers—both new consumers joining
the market and consumers from its competitors. The result is an
increased market size for the largest company in the market and
reduced market size for the rest. This phenomenon is called organic
growth (Locket et al., 2011)
. ◘ Figure 14.1 exemplifies this for
Facebook, in which a growth from 360 million MAU (monthly active
users) in 2009 to 608 million MAU in 2010 due to users churning from
competing social media and new users that have not used a social
media previously join the service. In this period (2009–2010),
Facebook performed few acquisitions (only one in 2009 and ten in
2010), and most of them were of technical nature. Hence, the growth in
users came only from either churning users or new users. Note that ◘
Fig. 14.1 presents the net increase in users—in the period some users
may have left Facebook for competing social media and vice versa.
However, the net flow of MAU from 2009 to 2010 was positive. Also
note that there are no or few detailed data whether the increase in
Facebook users was from churning or new users—only the net increase
is reported in official statistics. Most probably, the complete picture is
complex and involves churning across every social media at the time,
new users joining, and existing users quitting social media altogether.

Fig. 14.1 Facebook organic growth from 2009 to 2010. (Authors’ own figure)

Definition 14.1 Organic Growth

Organic growth is growth generated by the company’s own assets
and not from growth that comes from acquiring other companies.
The organic growth may be positive or negative.

However, there is another way for companies to grow, namely, through

mergers
and acquisitions (M&A
).
 Definition 14.2 Mergers and Acquisitions
A merger is a legal consolidation of two enterprises into one where
both enterprises cooperate equally in the merger, whereas an
acquisition implies that one company (the acquirer) takes control of
another company (the target company) by ownership of stocks,
assets, or equities of that company. Acquisition is also referred to as
takeover (Bannock et al., 1998).

In commercial terms, merger and acquisition are often treated as

synonymous because both result in a consolidation of two companies.
This is exemplified in ◘ Fig. 14.2 for the 2014 Facebook acquisition of
WhatsApp (Deutsch, 2020). The result is a company with joint assets,
employees, and customer base from the merged or acquired companies.
Facebook acquired WhatsApp—a competing social media with about
500 million MAU—for $19 Billion. It is one of the largest acquisitions
ever performed and has changed the business landscape in the social
media market toward increased concentration (less competition).

Fig. 14.2 The 2014 Facebook acquisition of WhatsApp. (Authors’ own figure)

Acquisition or takeover can be divided into four categories

depending upon how the acquisition takes place: friendly, hostile,
reverse, or backflip.
– Friendly takeover means that the takeover is agreed by the
management and the stockholders of both companies. However, the
takeover is initiated by one of the companies making it technically
different from a merger. In practice, there is no difference between
mergers and friendly acquisitions. The Facebook acquisition of
WhatsApp is an example of a friendly takeover (Deutsch, 2020).
– Hostile takeover implies that the management of the target company
is unwilling to accept the takeover, but for some reason, the takeover
nevertheless takes place. Methods may include offering a price well
over the market value of the company, persuading enough
stockholders to vote for a new and friendly management, or buying
stocks directly or via intermediaries to gain control of the company.
See, for example, Oracle’s acquisition of PeopleSoft (Daines et al.,
2006).
– Reverse takeover is a term used for takeover of a public company by a
private company so that the private company can go public
bypassing the complex process of going public itself. One example
from digital economy is the reverse takeover of the public US mobile
operator Metro PLC by the privately owned T-Mobile USA (Deutsche
Telecom) to form the new public company T-Mobile US, the third
largest mobile operator in the USA (Garza, 2013).
– Backflip takeover is an acquisition where the acquirer makes itself a
subsidiary of the acquired company, usually because the brand name
of the acquired company is better known. The Danish video game
producer Interceptor Entertainment acquired the American
company 3D Realms in 2014; moved its headquarters to Aalborg,
Denmark; and continued its business under the name 3D Realms
(Handrahan, 2014).

14.2 Motivations for M&A

Mergers and acquisitions are complex and expensive undertakings.
Hence, the involved companies must have clear business motivations
when initiating M&A. The overall motivation is that the outcome will
benefit business operations for the involved companies compared to
the situation before the M&A. This may include one or several of the
following:
– Getting rid of a competitor by first acquiring and then shutting down
the competitor; also referred to as killer or zombie acquisition
(Argentesi et al., 2020).
– Increasing the user base by merging the markets of the two
companies. The motive for a mobile network operator to acquire a
mobile network operator in another geographical region is to
increase the user base and, as result, increase both revenues and
value on the stock exchange.
– Achieving economy of scale advantages since merging two
companies often reduces common costs for administration, research,
marketing, inventory, and other expenses.
– Getting control of larger parts of the production and delivery chain;
for example, a mobile network operator acquiring retailers selling
smartphones, tablets, and other terminal equipment.
– Increasing market shares, for example, by attaching a competitor as a
subsidiary.
– Increasing revenues by absorbing a competitor and use the increased
market power to set prices and thereby raising the revenues to more
than the previous sum of revenues of the two companies.
– Improving geographical diversification by acquiring similar
companies offering the same service or good in a different
geographical region.
– Increasing the product portfolio, for example, offering video games as
a supplement on a social media platform.
– Utilizing synergy between different product categories, for example,
acquiring companies producing complementary products.
– Acquiring new skills and technologies, for example, buying promising
startup companies or manufacturers of highly specialized equipment
or services.
– Expanding into new profitable market segments. This is one of the
reasons why Google bought YouTube.
– Marketing under a more recognized brand. This is often the motive
behind backflip takeovers as explained above.
– Developing a promising concept, for example, by buying startups.
– Acquiring access to patents, protected content, shielded brand
names, and other IPR.
In the digital economy, there have been particularly many M&A as
shown in ◘ Table 14.1. Many of these companies have expanded into
almost all areas of ICT, becoming digital conglomerates. These
companies are constantly scanning the market for potential
acquisitions to increase the value of their business operations.
Table 14.1 Mergers and acquisitions in the digital economy. (Authors’ compilation)

Company Number of acquisitions Notable acquisitions

(by the end of 2020)
Google 240 YouTube for $1.65 billion in 2006
Microsoft 225 LinkedIn for $26 billion in 2016
Apple 121 Beats electronics for $3 billion in 2014
Amazon 103 Whole foods market for $13.7 billion in 2017
Facebook 89 WhatsApp for $19 billion in 2014
eBay 62 Skype for $2.6 billion in 2005

An examination of the acquisitions carried out by Amazon,

Facebook, Google, Microsoft, and Apple from 2015 to 2017 (a total of
175 acquisitions) indicates that the major motive has been to
strengthen the core business of the company and, more rarely, to
expand into new markets. Quite often the acquiring company shuts
down the business of the acquired company, even if it is not a direct
competitor, so that the motive apparently is to get access to core assets
such as skills, technologies, processes, practices, and IPR. Sometimes,
acquisition of startups seems to be substitutes for R&D. By acquiring
promising startups, the project gets proper funding and can be realized
in shorter time and may gain first-to-market advantages (Gautier &
Lemesch, 2020).

14.3 Types of Integration

Horizontal integration and vertical integration are two types of M&A.
 Definition 14.3 Horizontal Integration
Horizontal integration implies that a company merges with or
acquires another company in the same market segment.

The motive is either to get rid of a competitor or to build a company

with a larger customer base and increased economic value. Facebook’s
acquisition of WhatsApp in 2014 is an example of a horizontal
acquisition.

Definition 14.4 Vertical Integration

The motive of vertical integration
is to capture or secure a larger part
of the company’s supply chain. It is convenient to distinguish
between two types of vertical integration:
– Backward or upstream vertical integration implies that the
company acquires control over suppliers producing input to the
company’s own product; for example, an application service
provider merging with a content provider.
– Forward or downstream vertical integration
implies that the
acquires control over the delivery of services to the customers; for
example, a network provider (NP) offering direct access to the
users—that is, integrating with an ISP.

Google’s acquisition of parts of HTC in 2017 is an example of backward

vertical integration since Google took control over the production of
mobile phones (user equipment). Forward or downstream vertical
integration implies that the company acquires control over parts of its
delivery chain or enters related business domains, in which its prime
services are used as an input. Examples of forward integration are big
film studios (e.g., MGM) owning their own theaters and media
companies merging with Internet service providers and TV satellite
operators.
It is easier for the authorities to check that horizontal acquisitions
conform to antitrust laws and other market regulations than vertical
acquisitions. However, there is one particular problem associated with
regulations also in this case, namely, that many of the companies
acquired by the large companies are often young and small, so that the
acquisition is not regarded as a threat against healthy market evolution.
Due to the high number of M&A by companies in the digital
economy, it is important to analyze market structures and, in particular,
how these M&As may change competition. The role of national
regulators and other governing bodies is to supervise the market
evolution and to avoid the formation of de facto monopolies. The
distribution of market shares and revenues is not the only factor
determining whether a company will become a monopoly. The rate of
growth of the number of users or market share must also be
considered. If one company gets the lead in a market with strong
network effects, then with time, the company may become a de facto
monopoly.
Case Study: Facebook Acquisitions
Facebook has acquired 89 companies since 2007. It has moved from
being a social media service to a digital conglomerate with business
operations in several areas of the digital economy. A few notable
companies that Facebook has acquired include Instagram,
WhatsApp, Oculus VR, and Parse. ◘ Figure 14.3 exemplifies these
four acquisitions in the context of the types of integrations defined in
► Sect. 14.3:
– Horizontal integration: Facebook’s acquisition of Instagram in
2012. Instagram is a photo-sharing app launched in 2010 and
reached more than ten million users within a year. The intention
was that Facebook would develop Instagram independently;
however, integration with Facebook began shortly after the
acquisition.
– Horizontal integration: Facebook’s acquisition of WhatsApp in
2014. WhatsApp is offering messaging and voice-over-IP services
for its users. At the time of the acquisition, WhatsApp had more
than 500 million MAU. Even though Facebook and WhatsApp are
not competing directly in the same market, they are both
considered a part of the wider social media market. The
acquisition—together with the Instagram acquisition—has
increased Facebook’s influence and market power in the social
media market.
– Forward vertical integration
: Facebook’s acquisition of Oculus VR
in 2014. Oculus VR was developing virtual reality headsets. Such
devices may extend the reach of Facebook, by integrating the
social media to other types of devices.
– Backward vertical integration: Facebook’s acquisition of Parse in
2013. Parse was a mobile backend service provider. It gave
Facebook new tools for developing and accessing back-end
systems such as data storage, login management, and push
notifications.
 Fig. 14.3 Facebook acquisitions of Instagram, WhatsApp, Oculus VR, and Parse. (Authors’
own figure)

14.4 Conclusions
Mergers and acquisitions are common in the digital economy in which
the frequent motives have been to get rid of competitors, to get access
to new technologies, to buy assets such as patents, to increase market
shares, and to expand into new markets, or in short, to get bigger, more
powerful, and richer. Several of the big corporations are huge
conglomerates consisting of subsidiaries operating in different
technological branches. One reason being that the rate of innovations
after 1995, when the World Wide Web was commercialized, has been
enormous. Mobile apps are examples of this innovativeness. Currently,
there are almost three million apps available, and new apps are added
to the app store at a rapid rate, while old once that are no longer useful
disappear.
 All the big corporations in the digital economy have either become
de facto monopolies or are market leaders in at least one of their
markets. They have become so, because there are strong network
externalities in these markets that eventually work in favor of the
winner and keeps the competitors small or push them out of the
market. This is unavoidable organic growth. In addition, these
companies grow more by buying or merging with other companies,
creating commercial giants with a monopolistic core business
surrounded by a large number of subsidiaries. As has been
demonstrated in previous chapters, the company may not even earn
money from its core business but only from its subsidiaries. Facebook is
a good example.
This state of affairs is hard to regulate by the authorities to avoid
market failure and concentration of power.

Exercise
Google has made more than 240 acquisitions since 2001. A selection
of those include Android, YouTube, DoubleClick, GrandCentral,
Motorola, and Waze. Discuss how these acquisitions have allowed for
new or strengthened existing business domains of Google (you may
use Wikipedia for a description of Google and the acquisitions).

Answers
The key businesses of Google are spread over several technologies:
search engine, consumer services (email, software, hardware),
advertisements, and enterprise services. Here follows a few of them
and how they impacted Google’s business operations:
1.
Android Inc. was a company developing mobile phone software  
and operating system. It was acquired by Google to strengthen
its own efforts to develop software for mobile phones, e.g., for
supporting apps. Google created the Open Automotive Alliance
(OAA), comprising the world’s leading car manufacturers, to
promote the use of Android in cars. Google improved its position
as software developer by expanding the technologies of Android
Inc.
 2. YouTube pushed Google into the social media business. Google
has, by this acquisition, become an important factor in the social
media business and expanded its business area.  
3.
DoubleClick offered tools to advertisement agencies and media  
for increasing the efficiency of advertising by combining it with
the search engine technology. DoubleClick was merged into the
marketing platform of Google and Google Analytics to improve
their own algorithms for targeting advertisements.
4.
GrandCentral offered call forwarding and voicemail services.  
Google changed the name to Google Voice and, thereby, made its
entrance into the market for voice services (e.g., VoIP and
voicemail). In this case, Google bought the technology of
GrandCentral.
5.
Motorola Mobility produced smartphones, Bluetooth devices,  
cordless phones, set-top boxes, and other mobile devices. Google
acquired the company to get control over its patent portfolio
and, by this action, protect the Android operating system against
infringements.
6.
Waze is now a subsidiary of Google developing GPS navigation  
software. Before acquisition, Waze was a competitor to Google
Maps. The acquisition therefore has strengthened the position of
Google in GPS-related applications. For this reason, the
acquisition is still controversial.

References
Argentesi, E., Buccirossi, P., Calvano, E., Duso, T., Marazzo, A., & Nava, S. (2020, March 4). Tech-
over: Mergers and merger policy in digital markets. VoxEU & CEPR Policy Portal.

Bannock, G., Baxter, R. E., & Davis, E. (1998). The penguin dictionary of economics. Penguin
Books.

Daines, R. M., Nair, V. B., & Drabkin, D. (2006). Oracle’s hostile takeover of PeopleSoft (A). Harvard
Business Publishing: Education.
Deutsch, A. L. (2020). WhatsApp: The best Facebook purchase ever? Investopedia.

Garza, L. M. (2013, April 24). Metro PRC shareholders approve merger with T-Mobile USA.
Reuter.

Gautier, A., & Lemesch, J. (2020, January). Mergers in the digital economy. CESifo Working Paper
Series 8056, CESifo.

Handrahan, M. (2014, March 3). Interceptor Entertainment acquires 3D Realms.

Gamesindustry.​biz

Locket, A., Wiklund, J., Davidsson, P., & Sourafel, G. (2011). Organic and acquisitive growth: Re-
examination, testing and enlarging Penrose’s growth theory. Journal of Management Studies,
48(1), 48–74.
[Crossref]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_15

15. Standards
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Standards organizations – Open standards – Backward

compatibility

Learning Objectives
After completing this chapter, you should be able to:
– Explain the necessity of developing standards for systems,
protocols, and other functionalities of the information and
communication technology to ensure interoperability and
interconnectivity between devices and systems.
– Provide examples of organizations responsible for developing
different types of standards and identify the organization that
is likely to have specified a standard for a particular
technology.
– Analyze the impacts of standards on the market for digital
services, and, in particular, why standards commoditize
technologies and services.

15.1 Why Standards?

Standards are necessary to ensure interoperability between users and
providers, consistency of services, protection against abuse, and
maintainability of quality. In the information and communication
technology (ICT), standards are critical since the ICT infrastructure
consists of different types of technical equipment produced by different
manufacturers from different parts of the world (e.g., network
components, smartphones, and laptops). Lack of standards results in
situations where international deployment of services is not possible.
The need for standards in ICT is also evident from ◘ Fig. 15.1
showing that five interfaces need to be standardized to provide
interoperability:
1.
Application protocols between software modules (e.g., http  
between browsers and webpages, streaming protocols, and e-mail
protocols)
2.
Application programming interfaces (APIs) for programming,  
downloading, and program execution and runtime management
3.
Transport protocols (e.g., UDP/TCP/SCTP and the encryption  
protocol Transport Layer Security (TLS))
4.
Access protocols (e.g., Ethernet, Wi-Fi, and 4G)  
5.
Internet protocols (e.g., IPv6 and the encryption protocol IPsec)  
Fig. 15.1 Interfaces in the digital infrastructure. (Authors’ own figure)

That these standards are strictly followed is particularly important

for the development and sales of digital goods and services designed
for the global marketplace. Standards are needed for several reasons:
– Standards are the tool by which worldwide ICT markets can be
created, where the foremost prerequisite is that the devices at each
end of the connection are capable of communicating irrespective of
where they are located and to which ISP they are connected.
– Standards are required for creating competitive markets, for
example, for end user equipment. This includes procedures for how
to connect devices to the network, how different types of equipment
(e.g., laptops and smartphones) can interoperate, and how to locate
and identify remote equipment.
– Standards enable cooperation between stakeholders responsible for
performing different tasks in the execution of certain services, for
example, banking services where financial institutions may
cooperate with third party service providers for trusted customer
 identification (e.g., ID.me), authentication (e.g., mobile operators),
card verification (card issuers), and transaction managers (e.g., point
of sale operators). This requires not only technical standards but also
legal, economical, and managerial ones.
– Standards for distributed processing are required to allow
computers to cooperate in performing a common task where the
various elements of the service are executed at different computers
at remote locations. This category includes concepts such as cloud
computing and grid computing. Examples are the Internet itself,
massively multiplayer online games, air traffic control centers, and
large scientific simulation models requiring interconnection of
thousands of computers in a grid to become massive
supercomputers.
International ICT standards ensure interoperability between users
globally. These are standards related to telecommunications networks
(e.g., the Internet protocol stack and the family of mobile network
standards developed by 3GPP), presentation formats (e.g., HTML), and
Internet naming and addressing formats and usage (e.g., URL and URI).
These standards are not subject to legal agreements between countries;
that is, they are not de jure standards. On the contrary, the ICT
standards are de facto standards developed by manufacturers,
universities, voluntary groups, or individuals to support the
international ICT infrastructure. If the proposal is valuable, it may be
taken into use and thereby becoming an international standard. The
World Wide Web is the most evident example, starting as a project to
facilitate communications between CERN (the particle physics
laboratory outside Geneva, Switzerland) and cooperating universities
worldwide. Within a few years, the WWW became the de facto standard
for information posting and browsing on the Internet.
The layout and content of the digital good or service itself is not
usually subject to standards: there is no standard concerning the
content and presentation of, for example, Facebook, Twitter, Netflix, and
Apple and Android apps. To reach the market, they must only support
the standards required for accessing the Internet and for interactions
with the users (the World Wide Web standards).
 Definition 15.1 Standards
A de jure standard is obligatory, for example, the use of the radio
spectrum as defined in the Radio Regulations of the ITU.
A de facto standard is a convention, procedure, or technology
developed by the users of the standard (e.g., the 5G standard of
3GPP) or been selected by market forces (e.g., VHS as the standard
for videocassette recorders). The de facto standard is not obligatory.
A de facto standard is often referred to as industry standard.
An open standard is a publicly available standard (e.g., the
standards of ETSI, ITU, and 3GPP). The standard may be free of
charge or be subject to a usage fee or a patent licensing fee. Most de
facto standards are open standards.

15.2 Standards Organizations

There are several organizations and groups specifying and
standardizing ICT infrastructures, protocols, and operations. In this
section, we will just look at some of the organizations and groups
having the biggest impact on the evolution of the technologies and
services supported by ICT. Some of them are based on international
charters (ITU, ETSI, and 3GPP), while others are nonprofit interest
groups dedicated to a particular field of standardization (the Internet
Society and the World Wide Web Consortium). The standards produced
by these organizations and groups are open-source standards, meaning
that anyone may load down and apply the standards free of charge.
However, buying open-source standards from ITU and ETSI are often
subject to certain fees.

15.2.1 ITU
The world’s oldest standardization organization is the International
Telecommunications Union (ITU). ITU was established in 1865 for the
standardization of the emerging telegraph service. ITU was included as
a specialized organization in the United Nations in 1947. The union is
responsible for the standardization of telecommunications networks,
equipment, technical interfaces, network management, services, and
operations. This includes, in particular, the standards for the telephone
network and mobile networks and, to a lesser degree, the
standardization of the Internet and ICT. In fact, at the meeting of the
World Conference on International Telecommunications 2012 (WCIT-
12), the European Parliament presented a resolution where it “Believes
that the ITU, or any other single, centralized international institution
(e.g., ICANN), is not the appropriate body to assert regulatory authority
over the internet” (European Parliament resolution on the forthcoming
World Conference on International Telecommunications (WCIT-12) of
the International Telecommunication Union, and the possible
expansion of the scope of international telecommunication regulations
(2012/2881(RSP))). The major concern was that ITU regulations, in
particular on tariffing, may undermine the principle of network
neutrality. Several other countries supported this view, among others,
the USA, India, Australia, and Japan. Nevertheless, a new resolution was
accepted by 86 of 152 countries stating rather vaguely “to invite
Member States to elaborate on their respective positions on
international Internet-related technical, development and public-policy
issues within the mandate of ITU at various ITU forums including, inter
alia, the World Telecommunication/ICT Policy Forum, the Broadband
Commission for Digital Development and ITU study groups”
(International Telecommunications Union, 2012).
ITU is not the dominating organization behind the Internet today
and will most probably not be so in the future because of the opposition
expressed by the EU, USA, Japan, and several other technologically
advanced countries. For the evolution of the Internet and digital
services, the ITU may become an organization that is not generating the
standards, but rather ratifying standards produced by more specialized
organizations.
One of the most important tasks of ITU is to govern the use of the
radio spectrum. The allocations of radio spectrum to the different
services (satellites, land mobile networks, radio astronomy, radio
amateurs, broadcasting, and several other uses) are revised every 3 to 4
years by the World Radiocommunications Conferences (WRC). The
allocation of the frequency spectrum is an international de jure
standard. Other de jure standards of ITU include allocation of country
codes for telephone numbers, international mobile subscriber
identities (IMSI), and international numbering and identification plans
for radio communication with ships and aircraft.

15.2.2 ETSI
The European Telecommunications Standards Institute (ETSI) was
established in 1988 as an offspring of the Conférence européenne des
administrations des postes et des télécommunications (CEPT). ETSI is an
independent standardization organization for the EU and associated
European states (e.g., Switzerland, Norway, and Turkey). Industries and
organizations of these countries are the full members of ETSI. In
addition, there are several organizations and industries from other
counties outside Europe that are associated members, for example,
USA, Japan, People’s Republic of China, India, Brazil, Australia, and
Canada. Currently, ETSI has over 800 full and associated members
(countries, industries, and organizations).
ETSI is now regarded as world’s most influential, progressive, and
successful standardization organization on all aspects of information
and communications technologies, including new fields such as
machine-to-machine (M2M) technologies and the Internet of Things
(IoT). ETSI has taken over many of the roles ITU had previously,
publishing more than 2000 standards every year.

15.2.3 3GPP
The third Generation Partnership Project (3GPP) is responsible for
developing the standards for public land mobile networks 3G, 4G, 5G,
and beyond. 3GPP has also taken the leadership in developing Internet
standards for applications in mobile systems such as new voice-over-IP
standards, the IP multimedia subsystem (IMS) for application in all-IP
mobile systems, and access technologies and architectures for the
Internet of Things.
3GPP is a partnership of the major standardization organizations in
the USA, Europe, and Asia. The technical support team is located at the
headquarters of the European Telecommunications Standards Institute
(ETSI) in Sophia Antipolis, France. The standardization work is based
on voluntary contributions from more than 370 member organizations.
All standards made by 3GPP can be accessed and loaded down free
of charge by anyone and are, in this respect, free open-source standards
(The 3GPP specifications can be loaded down free of charge from ►
http://​www.​3gpp.​org/​specifications/​releases). One of the most
successful technological evolutions in ICT is the evolution of the digital
mobile telephone service, starting with implementation of GSM in
1991. The events leading to the standardization of mobile
communications are reviewed in ► Box 15.1.

Box 15.1 From GSM to 5G

Standardization of public mobile communications plays a
particularly important role in the evolution of digital services. The
successful standardization of GSM was also one of the major
arguments for establishing ETSI. Therefore, we will describe some of
the events leading to the current standards for mobile
communications.
The evolution of digital mobile communications started in 1982
when 17 European countries decided to jointly specify a pan-
European digital mobile network. The group set up for doing the task
was named Groupe Spécial Mobile, GSM. Later the system the group
specified was renamed the Global System for Mobile
Communications, also abbreviated GSM. In 1982, several
incompatible systems for land mobile systems existed or were about
to be put into operation in Europe: NMT in the Nordic countries, the
Netherlands, Switzerland, and Spain, TACS in the UK, C-Netz in
Germany, and Radiocom 2000 in France.
To ensure that GSM was built and not put aside as an interesting
future option, 13 European countries signed a memorandum of
understanding (MoU) in 1987 obliging that “operational networks
shall be procured in each of the countries by the network operators
based on the CEPT recommendations with the objective of providing
public commercial service during 1991” can be downloaded from ►
http://​www.​gsmhistory.​com/​wp-content/​uploads/​2013/​01/​5.​-
GSM-MoU.​pdf). Therefore, GSM operation could commence in Europe
in 1991/1992. GSM was not only built out in Europe; within a few
years, GSM had become the preferred mobile network standard in
most of the world.
GSM is a European standard that became a worldwide de facto
standard. What is more important is that the GSM standardization
 process became the norm by which all later mobile standards—3G,
4G, 5G, and variants thereof—are made. This includes features such
as service definition, network architecture, roaming, handover,
subscription module (SIM), addressing, and so on. The
standardization process is also an example of an open and dedicated
cooperation between companies that later would become
competitors as network operators, suppliers of network equipment,
and manufacturers of user terminals. This is a particular form of
coopetition. Coopetition implies that the companies may both
cooperate and compete either at the same time or at different stages
of the evolution. The reasons for coopetition in developing a
technological standard are several:
– Instead of one company or organization carrying the total
development cost, the cost is shared between several partners; the
total cost of developing the rather cheap GSM standard was more
than 100 million euros and required more than 1000 man-years of
expert work. The development of the 4G standard has required
several times as many resources.
– A global standard makes the total market pie much bigger, and,
consequently the market for each participant is also bigger.
– The economic risk of participation in projects based on standards
with global market potential is much smaller than for
implementing a local standard.
The work on a global mobile network standard was initiated in
ITU in 1986 under the name Future Public Land Mobile Network
System (FPLMNS). The work progressed very slowly, and no
significant results were obtained until 1998 when the project was
taken over by the newly formed organization third Generation
Partnership Project, 3GPP. Since then, this cooperation has developed
the 4G and 5G standards and is now expanding these standards to
support new services and features.

15.2.4 Internet Society

The Internet Society is an American nonprofit organization in charge of
promoting the standardization and policies of the Internet. The
organization also has several offices outside the USA (e.g., in Geneva
and Brussels to be close to both the UN policy group on information
technology issues and the political and technological power centers of
Europe). The organization has no legal influence on the ICT evolution.
On the other hand, the informal influence is enormous.
The Internet Society is the home for several legally informal
standardization bodies, the most important of which are:
– The Internet Engineering Task Force (IETF) is in charge of
developing and promoting Internet standards. There is no formal
membership of the organization, and anyone may contribute to the
work by issuing Requests for Comments (RFCs) which may contain
amendments or additions to existing standards or proposals for
completely new standards. The proposal may be accepted by the
Internet Engineering Steering Group and becoming a new Internet
standard. Even so, it may be rejected or ignored by manufacturers
and Internet providers and never be implemented. Being so loosely
organized, the Internet may evolve in an unplanned and haphazardly
way. This has so far been the major forte of the Internet.
– The Internet Architecture Board (IAB) is an informal advisory group
in charge of inducing some degree of consistency on the evolution of
the Internet, among others by sorting out RFCs that may become
useful additions to the Internet technology. This induces some
direction to the evolution of the Internet.
– The Internet Engineering Steering Group (IESG) is the forum that
finally endorses new Internet standards.
– The final Internet standards are also published as RFCs. Free access
for downloading of all RFCs is available via the homepage of the
Internet Society.
Note that ITU plays no important role in the standardization of the
Internet. On the other hand, 3GPP is playing a more and more
important role, in particular, in the development of the Internet of
Things.

15.2.5 World Wide Web Consortium

The World Wide Web Consortium (W3C) is an independent
organization in charge of standardizing web services. The organization
was established and is currently managed by the inventor of the World
Wide Web, Tim Berners-Lee. The charter is to standardize and develop
the WWW technology and promote WWW-derived services. This
includes presentation languages (XML, HTML), formats (XForms),
procedures (SOAP), and protocols (HTTP, HTTPS). The W3C standards
are independent of the Internet standards. The only requirement is that
the Internet exists as an underlying network for communications.

15.2.6 Institute of Electrical and Electronics Engineers

(IEEE)
Institute of Electrical and Electronics Engineers Standards Association
(IEEE-SA) develops standards within a broad range of technologies
where telecommunication is just one of them. The most important
standards are assembled in the 802-series. This series includes
standards for the Ethernet, Wi-Fi, Bluetooth, Zigbee, body area
networks, and other local area technologies. These technologies define
how various types of equipment can be connected to the Internet or
interconnected locally to form local area networks for different
purposes. While all the organizations listed above are authorized
standardization bodies either directly or through association with
other organizations, the IEEE is not. The IEEE is rather a loosely knitted
community of scientists and engineers participating in developing the
standards. Despite this, IEEE-SA is one of the most influential
standardization bodies in the world, having specified most of the local
communications technologies surrounding us.

15.3 Market Implications of Standards

Standards have significant implications on competition and on how
digital services evolve in the market. Standards are drivers of
commoditization—even complex services like mobile communication
and Internet access are commoditized. The user will, for example, not
experience any difference using smartphones from different
manufacturers or receiving the service from different mobile network
operators. Commoditized services compete primarily on price and not
on other features. This means that it is easy for users to switch to
competing service providers since all other features except price are
more or less the same. In such a market, it is difficult for the provider to
lock in consumers because the switching costs both for the consumers
and the supplier are small.
Other standards support diversity, for example, the standards of the
World Wide Web. These standards allow application service providers
to develop differentiated services satisfying various user needs. There
are also commoditized services on the web, for example, e-mail and
web browsing.
It is more likely that de facto monopolies develop in markets
without standards or with more than one competing standard because
a consumer must choose between equivalent services from different
suppliers that are technically incompatible. In this case, it is expensive
for the customer to switch to another supplier. Moreover, network
effects may dominate in the competition so that one of the providers
ends up as a monopoly. One example we have already encountered
several times is the competition between the video recording standards
VHS and Betamax in the 1970s and 1980s. VHS and Betamax offered
similar capabilities but were not compatible since there was no
common standard for video recording. VHS cassettes could not be
played on a Betamax recorder and vice versa. In fact, VHS and Betamax
were competing industrial standards developed by different
companies. Because of network effects, both standards could not
coexist in the market—over time one of the standards would
outcompete the other. By the mid-1980s, it became clear that—for
various reasons we will not discuss here—VHS had won this “videotape
format war.” All the engineering and marketing efforts put into Betamax
was in vain and had no benefit for the company developing it and for
society.
The narrative of VHS vs Betamax shows us that competition among
standards can be expensive. The lesson from this case is that it is better
for operators and manufacturers first to cooperate and thereafter to
compete once the standard has been agreed upon. This has been the
case for almost all ICT standards developed during the last 30 years.
Suppliers of equipment or services first cooperate to develop a
common standard. Once the standard is agreed upon, it is freely
available for any supplier of equipment or services, also those who did
not take part in the standardization process. The suppliers may then
develop products and services based on this standard and compete for
market shares. The benefit for the manufactures and the service
suppliers is that the cost to develop the standard is small for each
partner and that the total market becomes larger. This was the
successful approach taken by GSM and later by 3GPP in developing the
mobile standards.
There are also several examples where regional standards have
been developed, forming technological cartels. Europe and the USA
have often adopted different telecommunications standards, for
example, the basic encoding of voice signals in the digital telephone
network (PCM). These technological incompatibilities were solved by
interworking units at the borders converting the formats, thereby
satisfying the all-important requirement of interoperability.
Except for a few instances, GSM was not implemented in the USA, so
that roaming between Europe and the USA using the same mobile
phone was not possible. To respond to this inconvenience, several
mobile phone retailers saw the opportunity to establish a new
business, among others at airports, by leasing mobile phones to
travelers to Europe.
There were also two incompatible standards for 3G systems: UMTS
(specified by 3GPP) in Europe and CDMA2000 (standardized by ITU) in
North America and several countries in Asia; CDMA2000 was also
offered by several European mobile operators in parallel to UMTS to
enhance global roaming. The first true global standard for public
mobile networks was 4G specified by 3GPP and endorsed by ITU as a
global standard.

15.4 Standards and Interconnectivity,

Interoperability, and Backward Compatibility
Interoperability is the key feature of the Internet. Each user can
communicate with any other user or webpage—on e-mail or via web
browsers—independently of the technology employed by the other
user. Interoperability must therefore exist between networks designed
with different network technologies and between user equipment of
different brands and standards. Otherwise, the Internet will split up
into incompatible islands and loosing much of its value.
 In the physical network, technical standards and economical and
legal agreements are required to interconnect networks owned by
different ISPs. Nontechnical agreements may include remuneration for
transiting and terminating traffic, liabilities in case of network failure,
and terms of cooperation; technical agreements may include minimum
quality of service commitments such as availability, minimum
guaranteed data rate, and maximum latency and data jitter.
There are two incompatible network protocols on the Internet:
Internet Protocol version 4 (IPv4) and Internet Protocol version 6
(IPv6). IPv4 was specified in 1983 and is still used in several networks.
IPv6 was ready for implementation in 2006, but the adoption rate has
been slow until recently. IPv6 was developed to provide more
addressing space than IPv4. The adoption rate of IPv6 has now
increased rapidly because IPv6 is the only network protocol used in 4G
and 5G mobile networks. Moreover, the addressing space of IPv4 will be
too small for accommodating billions of connected IoT devices.
The Internet, therefore, consists of islands based on IPv4 and IPv6,
and a technology called tunneling is used to transfer IPv6 packets
across IPv4 networks and vice versa. Tunneling means that the IPv6
(IPv4) packets are imbedded in the data field of IPv4 (IPv6) packets. On
the other hand, most terminal equipment contains software for both
IPv4 and IPv6 so that the equipment can be connected to either type of
network. This ensures interconnectivity on the Internet.
Interoperability between mobile phones (smartphones) and mobile
networks is supported by backward compatibility, implying that, for
example, a 4G telephone can access 2G (GSM) and 3G networks. This is
achieved by implementing the radio and signaling interfaces of all three
standards in the telephone. This is possible since the evolution of
computers have followed Moore’s law: the processing and storage
capabilities of mobile phones have doubled approximately every 1.5
year. This means, for example, that the computational power of mobile
phones in 2001 when 3G was introduced was approximately 60 times
bigger than that of GSM phones of the same physical size. Similarly, the
computational power had increased by another factor of 60 when 4G
was introduced in 2010 and had increased by still another factor of 60
when 5G was introduced in 2019.
 The backward compatibility of mobile phones is achieved by
implementing the three standards in all phones and install algorithms
by which the phone can search for and identify the type of network
serving a particular area. Based on information displayed to the user,
the selection of network may then be automatic or manual. To assure
backward compatibility, the network operators operating a 4G network
must also operate, at least, a parallel 2G network. Several operators
plan to discontinue offering 3G networks since 4G offers much better
and faster Internet connections and because they still offer GSM
network access supporting earlier standards. However, there are
operators that also have shut down their GSM networks (e.g., USA and
Australia).

15.5 Standards for Trust and Security

Digital services often require cooperation between several
stakeholders. One example is banking. Such configurations require that
trust exists between the stakeholders and that trustworthiness can be
verified to a high degree of confidence. The trust relationships may
sometimes exist over several administrative domains (companies or
countries) with different legislations, rules of business conduct, and
regulations.
Trust may imply several things, for example (See the ISO/IEC 27000
family of standards for a detailed overview of recommendations on
information security and related procedures):
– Secure identification and authentication of communication partners
mean that the partners mutually verify the correctness of their
stated identities. Methods include permanent or onetime passwords
and cryptographic authentication methods. Secure identification
may include more complex procedures involving independent
trusted third parties.
– Non-repudiation implies that the originators and receivers of
information cannot deny their participation in the exchange of
information. This means that the supplier of the good cannot deny
having sent the electronic good, for example, deny responsibility if
the good contains malware that interfere with or damages the
computer of the receiver. Moreover, the supplier cannot deny having
 received payment for the good. On the other hand, the receiver of the
goods cannot deny having received the good, possibly including
encryption keys to decrypt the good. Non-repudiation may be
achieved by attaching digital signatures to the messages sent; for
example, attach the supplier’s digital signature to the good itself and
to encryption keys required for decoding encrypted goods and to
attach the receiver’s digital signature to messages acknowledging the
receipt of the good and associated encryption key.
– Certification implies that a trusted third party affirms the ownership
of certain cryptographic secrets such as keys used for digital
signatures, authentication, and encryption.
Trust is a legally complex issue. In many contexts, trust must be
based on legally binding covenants and be subject to criminal
proceedings if fraud is detected. Therefore, there are few, if any, trusted
third parties (TTPs) offering services outside small spheres of
influence, for example, specialized enterprises protecting interactions
between financial institutions and mobile network operators offering
two-step authentication for clients such as banks and governments.
Example of two-step authentication is cryptographic authentication of
the smartphone of the client followed by onetime passwords received
in SMS messages for authenticating the access attempt.
In the early years of the public Internet, it was expected that it
would be a lucrative business to be a trusted third party (TTP). Several
standards, for example, for public key infrastructures for secure
management of RSA encryption keys, were developed for this purpose.
The business potential was regarded to be huge, but all legal problems
and pitfalls associated with this business turned out to be many, and
the few attempts to establish such companies failed: no one would trust
the trusted party! TTPs owned by governments are not trusted because
the users of the TTP services may suspect that the government will use
the information collected by the TTP for clandestine purposes and
social control. Privately owned TTPs are not trusted because the
owners of the TTP may misuse the TTP for commercial reasons, for
example, interfering with the business of the user or selling
information gathered by the TTP to competitors. The TTP may also
represent a serious security threat because hackers may gain access to
the TTP tampering with or compromising the businesses of the users of
the TTP. The Dutch company DigiNotar issued certificates for
public/private keys for the Dutch government’s public key
infrastructure program. In 2011, hackers broke into the system and
issued fake certificates used for criminal purposes, for example,
attacking Iranian dissidents. The company went bankrupt in 2011 as a
result of the break-in (See the Wikipedia article about DigiNotar).

15.6 Conclusions
Interoperability is one of the key features of ICT. Interoperability means
that all networks (mobile or fixed) are interconnected and are able to
pass messages between people and machines irrespective of in which
country or region of the world they are located. Interoperability also
implies that equipment produced by different manufacturers can work
together using standardized interfaces and protocols.
International standards for ICT are developed by several specialized
standards organizations. Almost all ICT standards are de facto
standards; that is, they are not mandatory but are convenient since they
ensure global interconnectivity and support innovations of applications
and services that otherwise would have been impossible. Just a few
international ICT standards are de jure standards. Examples are the use
of the frequency spectrum by different radio communication services;
the formats of international identification and numbering plans for
land mobile, aeronautical, and maritime services; and standards for
certain services such as ground-to-air traffic control.
The standards also commoditize the technologies and services they
specify, for example, wireless communication, Internet access, and
World Wide Web protocols. On the other hand, these commoditized
technologies support a vast number of non-standard applications
opening up for a digital market consisting of a mix of millions of big and
small businesses. The three basic technologies—wireless
communication, Internet, and the World Wide Web—have created an
enormous, innovative arena for business development.

Questions
1. Which organizations are responsible for standardization and
allocation of:  
 (a)
Telephone numbers?  
(b)
E-mail addresses?  
(c)
Web addresses?  
Hint: this information is found by searching the
Internet.
2.
How can a 4G smartphone communicate with a GSM phone for  
voice communication despite being designed to incompatible
standards?
3.
Which protocols are standardized by RFC791, RFC2616, and  
RFC793?

Answers
1. The standardization bodies responsible for allocation of
telephone numbers, e-mail addresses, and web addresses are:  
(a)
ITU has standardized the general formats of international  
telephone numbers and is responsible for allocating
unique country codes identifying the country. This code
consists of the first one, two, or three digits of the
international telephone number. The remaining digits
(called the national telephone number) are allocated by
national authorities, either a regulatory authority or the
telephone network operator.
(b) The e-mail address is written as local-name@domain-
name (e.g., Joe.​jones@example.​edu) where domain name is  
accredited and registered by the Internet Corporation for
Assignment of Names and Numbers (ICANN), and the local
name is allocated by the e-mail provider. The allocation
principle is standardized by the Internet Engineering Task
Force (IETF). The domain name identifies the host and is
 converted to a unique IP number for sending the e-mail
over the Internet to the e-mail server.
(c)
The format of the web address was proposed by Tim  
Berners-Lee and standardized by IETF. The format is
composed of three parts as follows: ptotocol://► www.​
domain-name/​index.​html. The protocol is either http or
https, the domain name is the same as for e-mail addresses
(allocated by ICANN), and the index is the local address
generated automatically by the file management system of
the host computer. The index identifies the web page
uniquely within the file system of the host computer.
2.
The voice format is translated from the GSM format to the 4G  
format—and vice versa—by interworking units in the network.
The same applies to conversations between 4G (or GSM) and
fixed telephones, between ordinary telephones and VoIP
telephones, between different standards of fixed telephones
(e.g., between European and North American coding standards),
and so on.
3.
RFC791, IP version 4; RFC2616, http; and RFC793, TCP.  

Reference
International Telecommunications Union. (2012). Final acts: world conference on international
telecommunications, Dubai.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_16

16. The Long Tail

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Heavy tail – Power-law

Learning Objectives
After completing this chapter, you should be able to:
– Understand the concept of long tail distributions.
– Identify the existence of long tails in a particular digital
service.
– Evaluate the effect of the long tail on revenues.

16.1 Origin of the Concept: Amazon and the

Long Tail
The long tail refers to goods and services that are in low demand
individually but collectively constitute to substantial sales. The term
was coined by Chris Anderson in his 2006 book, The Long Tail: Why the
Future of Business is Selling Less of More (Anderson, 2006).
Traditionally, such goods and services have been too expensive for a
seller to offer to its customers. Let us take a bookstore as an example. A
physical bookstore can accommodate a limited number of books,
typically, between 10,000 and 100,000 books. This is because the
bookstore has finite space for storing and displaying books. A
bookstore offers and displays books that are believed to sell in large
quantities, thereby reducing storage costs and maximizing revenues.
Books believed to have few sales are not offered since the cost of
holding such books does not match the forecasted revenues they may
generate.
A digital bookstore does not have the same limitations as a
traditional bookstore. Millions of titles can be stored in a single server
and be available for anyone from anywhere. Bestsellers do not take up
more space than the rarest titles. The same applies to all types of digital
goods such as music, films, and video shows.
Amazon can offer millions of different books on its e-commerce
channel. This is achieved as a combination of large and well-organized
warehouses containing printed books, digitally stored books, on-
demand printing, third-party mediation, and personalized advertising
based on search and purchase history of the customers. The most
popular books sold by Amazon are printed in advance and stored
physically in large, efficient, and automated warehouses. These
warehouses may hold what is termed “the head” of Amazon’s products;
that is, the books that sell the most. Other books—both bestsellers and
books which are low in demand—are supplied by Amazon as e-books
or are printed on demand. E-books are stored digitally, and there is
almost no cost for Amazon to add an e-book to its inventory (again an
example of zero marginal cost). On-demand printing means that
Amazon offers physical books to customers by printing them when they
are ordered by the consumer. Amazon is also a multisided platform
(MSP) and mediates between third-party sellers (bookstores or
authors) and customers; that is, other bookstores and authors use
Amazon’s digital marketplace to offer their books for sale (Hanks,
2017). These bookstores may be niche stores offering books that are
low in demand. This arrangement expands Amazon’s supply of books.
Finally, Amazon may use the search and buying history of the
customers to suggest other books, including rare books, which may
interest the buyer.
The sum of these supply mechanisms results in Amazon being able
to offer millions of books to the customers. The bestsellers are sold by
Amazon directly by printing in advance or as e-books. Books that are
low in demand are offered through a combination of third-party
booksellers, e-books, and on-demand printing of digitally stored books.
Hence, Amazon can accommodate the 10,000–100,000 books offered
by a physical bookstore (the head) and, in addition, millions of other
books not offered by the physical bookstore. These millions of other
books that are not offered by physical bookstores constitute the tail—
books that are low in demand individually, but collectively constitute
substantial sales. This is called the long tail
of supply.
An illustration of the long tail concept is shown in ◘ Fig. 16.1. Here,
products are ranked (abscissa) according to sales (ordinate), where the
top-selling products are those in the head and the books that sell the
least are in the tail. Chris Anderson observed that companies like
Amazon earned about half its revenue from products in the tail.
Anderson based his conclusions on observations made by Brynjolfsson
et al. (2003). Brynjolfsson and coworkers found, for example, that on
Amazon, 2.3 million book titles were available, while the shelves of an
ordinary large bookstore contained between 40,000 and 100,000 titles.
They estimated that the sales of books not found in ordinary stores
amounted to between 20% and 40% of the total sales of Amazon. In a
new survey published in 2010, they found that 36.7% of Amazon’s sales
came from the long tail (Brynjolfsson et al., 2010). This is illustrated in
◘ Fig. 16.2.
Fig. 16.1 The long tail. (Authors’ own figure)

Fig. 16.2 Amazon sales. (Authors’ own figure)

16.2 Internet and the Long Tail

In a paper from 2011, Brynjolfsson et al. also argued that the long tail
phenomenon is a common aspect of many Internet businesses
(Brynjolfsson et al., 2011). Low production costs, cheap storage, small
shipping costs, and efficient information searches are the key
ingredients for long tail businesses. They emphasized the importance of
efficient information search both on the web and on homepages of
vendors such as Amazon, allowing customers to skim through millions
of products using search queries based on product names, key words,
product description, product category, or other information.
Offering products in the long tail is a competitive advantage for
providers. This is because customers prefer variety and options when
selecting products to buy. The long tail also gives the provider more
relevant options for advertising and recommending similar products
for the customers. Hence, companies that offer products in the long tail
are rewarded by both increased sales and improved customer
satisfaction by displaying a long list of products to choose from
(Brynjolfsson & Saunders, 2013).
The long tail has been exploited in many digital businesses. In the
music industry, the digitalization and streaming of music means that
more titles can be offered to customers compared to a traditional music
store. Spotify, for instance, offers access to more than 30 million songs,
while the biggest record stores accommodate around 100,000 albums
(equivalent to about one million songs). For Spotify, there is no cost
associated with storing one extra song and making it available to users
(zero marginal cost). Moreover, Spotify and other digital music provides
may also de-bundle albums and sell individual songs.

Definition 16.1 The Long Tail

The long tail refers to goods and services that are in low demand
individually but collectively contribute to substantial sales.

In the banking industry, the long tail has been exploited to provide
microcredit to lower-class and poor people. For instance, Grameen
Bank in Bangladesh offers small loans to people and private enterprises
that would normally not be qualified for a loan in a regular bank.
Crowdsourcing is an example of long tail production. The long tail
here is made up of all the skilled people that may contribute to a
software product but are, a priori, unknown to the developer. Wikipedia
encourages the regular Internet user to become an author of the
encyclopedia, hence creating a long tail of contributors. Wikipedia’s
workforce is supplied by many contributors in a long tail of supply.
Big companies, such as Amazon, can profit from all products—both
products that sell in huge numbers (head) and products that sell in
small numbers (tail). However, the long tail has also made it possible
for individuals to start a type of business that was not possible before.
For example, individuals may now sell merchandise on eBay, offer their
services on Airbnb and Uber, and sell their own books on Amazon.
These individuals supply the long tail with products and services that
were not previously available to consumers. Selling merchandise
without a physical store was expensive before the advent of e-
commerce. However, today, anyone can become an e-commerce retailer
(or reseller) by setting up a web page to offer manufactured products.
Such a business will find its place on the long tail among thousands of
other suppliers in the same business area.

16.3 Numerical Analysis of the Long Tail

The term “long tail” alludes to statistical distributions where a large
portion of occurrences are far away from the main part of the
distribution. In statistics, the long tail distribution is a special case of
heavy tail distributions defined as follows.

Definition 16.2 Heavy Tail Distribution and Power-Law

Distribution
The discrete heavy tail distribution is a statistical distribution in
which the tail of the distribution falls off more slowly than
exponential decay. Any distribution of the form f(X = k)~k−γ, where γ
> 1 is a constant and the positive integer k is the running variable, is
a discrete heavy tail distribution. This particular distribution is also
called a discrete power-law distribution (Schroeder, 2009). The
symbol f(X = k) stands for “the frequency of occurrences that X is
equal to k,” and the tilde (~) means “is proportional to.” If γ is small
(between 1 and 2), then it is also called a long tail distribution.
Anderson, based on the works by Brynjolfsson et al., found that the
popularity of books sold by Amazon seemed to follow a rank
distribution in accordance with Zipf’s law; that is, a long tail
distribution with γ = 1 and where the probability drops to zero for all k
larger than a certain upper threshold (the cutoff). In 2000, the number
of titles available on Amazon was 2.3 million, while ordinary
bookstores held, on average, about 40,000 titles. The “long tail” of
Amazon then consisted of more than 2.2 million books, that is, books
available on Amazon, but not in an ordinary bookstore.
Let us then apply Zipf’s law to the sale of books. The number of
titles held by Amazon is 2.3 million books, while an ordinary bookstore
holds about 40,000 titles. The titles held by an ordinary bookstore are,
according to Zipf’s law, the most popular titles, ranging from 1 to
40,000 in popularity. The titles with popularity ranging from 40,001 to
2,300,000 in popularity are the long tail. The relative number of sales of
books from the long tail is then:

Hence, the long tail makes up 26.6% of all sales of books by

Amazon, provided that the demand for books follows Zipf’s law. Note
that empirical results based on Amazon sales yielded a long tail of
36.7%. This means that Zipf’s law gives a reasonable estimate of
Amazon’s long tail sales under the rather arbitrary assumption that
40,000 titles are in the head of the distribution.
If the tail is twice as long (4.6 million books), then the sales from the
tail increase to 29.8%; that is, the sales from the first 2.3 million books
in the tail amounts to 26.6% of the total sale, while the next 2.3 million
books in the tail amounts to only 3.2% of the total sales. If the tail is
only half as long (1.1 million books), the sales from the tail are 22.9% of
the total sales.
► Box 16.2 contains a generalization of Ziff’s law applied to
infinitely long tails.
 ► Box 16.3 is an example of the application of Zipf’s law to a
completely different area, namely, the vulnerability of the Internet and
the World Wide Web to targeted attacks on the infrastructure. Both the
Internet and the Web may be regarded as a long tail network. On the
Internet, most of the routers are small and connected by
communications links to few other routers, while the “tail” of the
Internet consists of rather few big routers connected to thousands of
other routers. In the Web, most websites are connected to few other
websites. These websites make up the “head” of the Web. The “tail” of
the Web consists of websites connected to very many other websites,
where search engines are example of websites connected to an
enormous number of websites.

Box 16.1 Zipf’s Law

Zipf’s law is based on the observation that the most frequent word in
English (“the”) is twice as frequent as the second-most frequent
word (“of”), three times as frequent as the third-most frequent word
(“and”), and so on. Zipf’s law holds quite well for, at least, the first
1000 words in the English language (Schroeder, 2009). The
frequency of words is then derived from the harmonic series:

or more precisely, the frequency f(k; N) of the k-th-most frequent

word is:

in which N is the number of words in the English language and γ ≈

0.57722… is the Euler-Mascheroni constant. We have used the fact
that:
 The notation O(1/N) (the “big O” notation) indicates that this
term decreases at least as fast as 1/N as N increases. For large N, the
last term can, therefore, be ignored. The statistical distribution with
frequency f(k; N) is also called the Zipfian distribution
. Note that the
distribution depends on the cutoff N.
Zipf’s law describes, in addition to word usage, the rank
distribution of amazingly many natural and sociological phenomena:
size of cities, size of countries (except China and India), length of
rivers, size of sand grains, wealth among people, and, as we have just
seen, popularity of books.

Box 16.2 Generalization of Zipf’s Law

The size of the tail can be adjusted by applying a general discrete
dissemination instead of Zipf’s law. In a general discrete power-law
distribution, the frequency of exactly k events is:

in which is the Riemann zeta function of argument

a. This distribution is called the zeta distribution.
The parameter a alludes to the size of the tail of the distribution.
If the tail starts at k = K = 40,001 (as in the example with Amazon
above), then the relative number of books, R, sold in the tail will be:

The Zipfian distribution is the special case in which α = 1. In this

case, there must be an upper cutoff N for which f(X > N) = 0, and the
zeta function is replaced by . The same applies if α < 1 since

the zeta function diverges for α ≤ 1.

 ◘ Figure 16.3 shows the relative number of books sold in the tail
(R) as a function of α. Here, the cutoff value is N = 2,300,000,
corresponding to the number of books available on Amazon. There
are 40,000 books in the head of the distribution. Observe that the
relative number of books sold in the tail (R) decreases from 98.3%
when α = 0 to about 0% when α > 1.6. For α = 0, all book titles (both
in head and tail) sell in the same numbers. In this case, sales are
uniformly distributed. Furthermore, for α > 1.6, the tail is too small to
have any economic value since books in the tail collectively
contribute to virtually no sales. Note that in the Zipfian distribution α
= 1, resulting in R = 26.6% as calculated above. Note that α ≈ 0.94 will
match the empirical data of Amazon’s sales R = 36.7% for the
parameters N = 2,300,000 and k = K = 40,001. In other words,
Amazon book sales as presented here can be modeled accurately
using general discrete power-law distribution with α = 0.94.
Fig. 16.3 Relative number of books in the tail as a function of alpha. (Authors’ own figure)

Box 16.3 The Long Tail and the Vulnerability of Internet

Power-law distributions are not only important to determine the
economic value of the long tail. These distributions are also used to
evaluate the robustness and vulnerability of the Internet and digital
services. In 1999, the two physicists, Albert-László Barabási and Réka
Albert, discovered that the number of hyperlinks pointing into or out
of webpages followed a general power-law distribution with an
exponent approximately equal to 2 (Albert & Barabási, 2001). The
same researchers also developed a general theory in which they
demonstrated that the number of links connected to the nodes of a
graph follows a power-law distribution if the graph is grown with
 preference. For the Web, “growth with preference” means that a new
webpage is connected to another webpage with a probability
proportional to the number of existing connections to that webpage.
This phenomenon is called the “Barabási-Albert (BA) random graph
model.” Later, it was found that the size of Internet routers measured
in terms of the number of connections they have with other routers
also follows the same power law (Dorogovtsev & Mendes, 2001).
Hence, the Internet and the Web have long tails. In these cases,
the tail consists of the largest and the head consists of the smallest
routers and webpages. This insight directs us toward another
observation, namely, that structures like the Internet and the Web are
very vulnerable to attacks against routers and webpages in the long
tail. Remove several of the biggest Internet routers—there are not
very many of them—and the Internet runs into severe connectivity
and capacity problems; remove the search engines from the Web, and
the Web falls apart. The Internet and the Web are vulnerable to
targeted attacks. On the other hand, most of the routers and
webpages have small connectivity, and a random attack on routers
and webpages may have little overall effect on the Internet or the
Web. Hence, the Internet and the Web are, at the same time, both
vulnerable and robust against failures and cyberattacks (Audestad,
2007).

16.4 Conclusions
One of the most important characteristics of digital businesses is that
the marginal costs of the products are zero; that is, there is no cost
associated with production, storing, and distributing the good. For
companies selling digital goods, it is, therefore, no practical limits to
how many products they have in store. This has created a new type of
business that is not practical for products requiring physical space for
manufacturing, storing, and delivery. This is the long tail of products
that are low in demand but collectively create considerable revenues.
Amazon has successfully applied this strategy, and it is claimed that
about one-third of their revenues stems from the long tail consisting of
books not available in ordinary bookstores. There are several examples
of businesses in the digital economy exploiting the same strategy.
 Statistically the long tail distribution is related to other statistical
distributions such as discrete power-law distributions, the discrete zeta
distribution, the Zipfian distribution, and general discrete heavy tail
distributions. For example, the long tail distribution of Amazon can be
modeled using the Zipfian distribution. Using the distribution fitting
the empirical data best, the expected sales from the long tail can be
estimated for a variety of digital businesses so that the concept can also
be used in strategic business planning.
The distribution of the size of routers on the Internet and
connectivity of webpages of the World Wide Web have also long tails.
This has important consequences both for the robustness and the
vulnerabilities of these structures.

Questions
1.
How is Airbnb exploiting the long tail of demand and the long  
tail of supply in its business operations?
2.
What are the characteristics of the long tail generated by Uber in  
the personal transport industry?
3.
What is the probability that that the Internet contains a router  
with 1000 connections if the size distribution of Internet routers
follows a general power law with α = 2? To simplify calculations,
you can use the fact that Riemann’s zeta function with argument
2 is . If there are ten million routers on the Internet,

how many are expected to have 1000 connections? How many

routers have exactly one connection?

Answers
1. Airbnb allows people to rent out their homes. This is how
Airbnb thereby creates a “long tail” to the supply of available  
vacation resorts for tourists. Airbnb contributes to the supply
side of the tourist industry. Airbnb are different from hotels.
They have few rooms available (typically accommodating 2–5
 guests) and may be situated outside the city center, where most
hotels are located. They may also be cheaper or offer facilities
not available at hotels. The demand aspect of this is that people
often prefer to rent unique homes, vacation homes, or rooms
with bread and breakfast rather than hotels.
2.
On the supply side, Uber creates a long tail in the taxi market by  
enabling private drivers to offer transport using their private
cars. This includes that the driver may offer special service such
as a ride in a sport car, limousine, van, motorcycle, and so on. On
the demand side, the market for personal transport has been
increased both in size and diversity.
3.
Apply the general power-law distribution and observe that:  

If there are ten million routers on the Internet, it is expected that

107 × 6.08 × 10−7 ≈ 6 routers have 1000 connections. The expected
number of routers with one connection is:

References
Albert, R., & Barabási, A.-L.. (2001). Statistical mechanics of complex networks. ArXiv.

Anderson, C. (2006). The long tail: Why the future of business is selling less of more. Hyperion.

Audestad, J. A. (2007). Internet as a multiple graph structure: The role of the transport layer.
Information Security Technical Report, 12(1), 16–23.
[Crossref]

Brynjolfsson, E., & Saunders, A. (2013). Wired for innovation. The MIT Press.

Brynjolfsson, E., Hu, Y., & Simester, D. (2003). Consumer surplus in the digital economy:
Estimating the value of increased product variety at online booksellers. Management Science,
49, 1580–1596.
Brynjolfsson, E., Hu, Y., & Smith, M. D. (2010). The long tail, the changing shape of Amazon’s
sales distribution curve. SSRN.

Brynjolfsson, E., Hu, Y., & Simester, D. (2011). Hello long tail: The effect of search cost on the
concentration of product sales. SSRN.

Dorogovtsev, S. N., & Mendes, J. F. F. (2001). Evolution of networks: From biological nets to the
internet and WWW. Oxford University Press.

Hanks, J. (2017, April 4). Amazon doesn’t do long tail. Why should you? Practical Ecommerce.

Schroeder, M. (2009). Fractals, chaos, power laws: Minutes from an infinite paradise. Dover.

Further Reading
Anderson, C. (2006). The long tail: Why the future of business is selling less of more. Hyperion.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_17

17. Digital Markets
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords E-commerce – Network access – Information services –

OTT – Pro sumer

Learning Objectives
After completing this chapter, you should be able to:
– Understand digital and e-commerce markets.
– Identify the different types of stakeholders in digital markets.
– Analyze network access markets and information service
markets.

17.1 Market Types

A market is a mechanism for trading both tangible and intangible goods
and services. ◘ Figure 17.1 classifies markets according to the type of
good or service (horizontal axis) and type of channel used for trading
the good or service (vertical axis). Examples of tangible goods include
computers and cars. Examples of non-digital services are hairdressing
and taxi rides. Online trading implies that an ICT infrastructure (the
Internet) is used to carry out some or all activities associated with the
trade: viewing products, bargaining prices and delivery terms,
ordering, product delivery, and transfer of payment. Tangible goods and
non-digital services can be traded online; however, these products
cannot be delivered over the ICT infrastructure. Shipment and delivery
of digital services is done online. We may then define a digital market as
follows.

Fig. 17.1 Classification of markets. (Authors’ own figure)

Definition 17.1 Digital Market

A digital market is a mechanism for online trading of both digital
goods and digital services and tangible goods and non-digital
services.

E-commerce is online trading of all kinds of goods and services. This

includes online shopping, online payment, transfer of funds,
management of supply chains, and business-to-business exchange of
data. E-commerce also includes the trading of network access and
information services. To classify an activity as e-commerce, it must
support some sort of digital payment system.
 E-commerce markets have grown to constitute an important part of
the global economy following the commercial success of the World
Wide Web (WWW) in the early 1990s. Two important and distinct parts
of e-commerce are the online markets for network access and the
online markets for information services:
– E-commerce markets are the business of selling goods and services
over the Internet (see ► Sect. 17.3).
– Network access markets are the business of providing access to the
Internet and other communication networks (see ► Sect. 17.4).
– Information service markets are the trade of content, applications,
and information on the Internet (see ► Sect. 17.5).
Note that not all activity performed online in digital markets is e-
commerce. For instance, information services may be exchanged
between provider and consumer free of charge. Such an exchange of
digital goods is not e-commerce since there is no payment involved.

17.2 Stakeholders and Relationships in Digital

Markets
◘ Figure 17.2 shows the most important stakeholders involved in
digital markets and the relationships between them. The network
provider (NP)
is the owner of the ICT infrastructure needed for online
trading, encompassing fixed networks, mobile networks, Internet
infrastructure, and storing and computing facilities. An infrastructure
provider owning a mobile network is sometimes referred to as a
“coverage” operator. The Internet service provider
(ISP) buys access to
this infrastructure from the NP and resells it to the consumer (C) and
the application service provider
(ASP). The consumer may use these
services directly to access the Internet, make phone calls, and send
SMS. The ASP uses the infrastructure access purchased from the ISP to
support the distribution of content, applications, and services that the
ASP produces. The ASP may also buy copyrighted content from a
content provider
(CP) such as movies, music, and news articles. The ASP
uses the input from the ISP and the content provider to offer digital
services and applications to the consumer.
Fig. 17.2 Stakeholders and relationships in digital markets. (Authors’ own figure)

Examples of types of services offered by ASPs are online music

streaming (e.g., Spotify and Tidal), online video streaming (e.g., Netflix
and HBO), digital newspapers (e.g., The New York Times
and Financial
Times), online banking (e.g., HSBC and Nordea), cloud storage (e.g.,
Dropbox and Google Drive), and social media services (e.g., Twitter and
Facebook). Many of the applications and services offered by the ASP are
free of charge for the consumer, for example, Google’s search engine,
Wikipedia, and Facebook. The ASP providing these applications must
acquire revenues from other sources than the consumers. Normally, the
consumer pays for Internet access to the ISP. Network access markets
are the business domain of the ISP and the NP, while markets for
information services are the business domain of the ASP and the CP.
The model in ◘ Fig. 17.2 is a simplification of the business domain
for digital services. The most important observation from this simple
model is that there is a sharp separation of business domains in the
provision of digital services. Not only are there many stakeholders
involved in the provision of digital services, but there are also big
differences in how each of them conducts their business. For instance,
the NP builds and operates a physical ICT infrastructure consisting of
optical fibers, mobile base stations, communication satellites, undersea
cables, Internet routers, and switching centers. This infrastructure may
cover a large geographical area and consist of expensive equipment.
Moreover, the NP needs a staff of technicians and engineers to build and
manage the ICT infrastructure. There are also huge undertakings in
upgrading the ICT infrastructure since ICT tends to get outdated
quickly.
The content provider produces content such as music, movies, and
news articles. The business operation of the content producer is vastly
different from that of the NP. The products of the content provider are
usually digital and reside only on a digital storage device. These
products do not need upgrades or extensive management after being
produced.
Formal relationships or contracts, in terms of service-level
agreements (SLAs), may exist between the stakeholders in digital
markets.

Definition 17.2 Service-Level Agreement

A service-level agreement (SLA) is a contract that exists between a
consumer and a provider of a digital service; see ◘ Fig. 17.3
. The
SLA describes certain contractual terms related to the delivered
service, e.g., price, delivery precision, and responsibility.

Fig. 17.3 Service-level agreement configuration. (Authors’ own figure)

Most SLAs between consumers and providers are standardized. If the
terms defined in the SLA are not satisfied by either party, some form of
compensation might be requested.
Examples of the terms that may be covered in an SLA include
specifications of maximum service delays, mean time between failures,
service availability, security, and privacy. All these specifications
describe the quality of the service (QoS) as perceived by the consumer.
Examples of SLAs are the contract between consumers and Internet
service providers for the delivery of network access, the contract
between consumers and Spotify for the delivery of online music, and
the contract between a provider of security alarm services and a mobile
network operator.

17.3 E-Commerce Markets

E-commerce is the online trading of tangible goods, digital goods, and
services. This is illustrated in ◘ Fig. 17.4. For all types of e-commerce
trade, the consumer conducts and manages the trade using an online
channel such as the Internet. The supplier handles the trade and ships
the products or services to the consumer. Digital goods and services are
delivered to the consumer over the Internet, while tangible goods are
delivered to the consumer using traditional transportation.
Fig. 17.4 E-commerce trading. (Authors’ own figure)

Examples of e-commerce trading include buying books from

Amazon, buying music from iTunes, purchasing electronics on eBay,
and subscribing to services delivered by Spotify, Netflix, mobile
network operators, and Internet access providers. In fact, all trading
activities conducted online can be categorized as e-commerce.
However, to categorize an exchange of goods and services between a
supplier and a consumer as e-commerce, there must be some sort of
financial activity between them. The exchange of goods and services
without any financial activity is not regarded as e-commerce even
though the trade is done in a digital market. One example is the use of
Facebook. The use of Facebook is free of charge for the user, and,
therefore, there is no financial activity between the user and Facebook.
The access to and use of Facebook is not e-commerce. On the other
hand, Facebook sells advertisement space to retailers and other
companies as a part of their business model. This is, indeed, e-
commerce in which Facebook is the supplier.

Definition 17.3 E-Commerce

E-commerce is the online trading of tangible goods and digital goods
and services with some sort of financial activity.

One particular form of e-commerce is mobile commerce, or m-

commerce, where all interactions between retailer and customer
(viewing, ordering, shopping, and payment) take place via smartphone
applications. The term was coined by Kevin Duffey in 1997. He defined
m-commerce as “the delivery of electronic commerce capabilities
directly into the consumer’s hand, anywhere, via wireless technology”
(Global Mobile Commerce Forum, 1997). The use of smartphones for
shopping and payment services is now taking over most of the e-
commerce market because of the convenience of using a smartphone
for this purpose rather than personal computers (Popovic, 2019).
In the USA, about 10% of all retail is performed using e-commerce
(E-commerce in the United States, Statistics and Facts, 2020)
. In China,
which is the largest e-commerce market in the world, about 20% of all
retail sales are e-commerce (Long, 2017). These numbers are from
2017, and they are expected to escalate in the near future. Similar
trends are seen in most parts of the world—the share of e-commerce is
increasing and is replacing traditional retail. Ecommerce News Europe
reports that e-commerce increased dramatically during the COVID-19
lockdown (The impact of Covid-19 on ecommerce, 2021). According to
the survey, 90% of the major companies involved in e-commerce
increased their sales during this period. For 50% of them, the sales
more than doubled.
Important milestones in the evolution of e-commerce were the
launch of eBay and Amazon in 1995, PayPal in 1998, and Alibaba in
1999. Amazon is now among the five largest companies worldwide
according to market cap. PayPal was one of the pioneers of online
payment systems. Other important services and companies in the e-
commerce market include Groupon (launched in 2010), Apple Pay
(launched in 2014), and Google Pay (launched in 2015 as Android Pay).
 There are two important requirements for successful e-commerce
markets.
– To become an efficient marketplace, e-commerce requires websites or
apps
where vendors can present their items for sale and buyers can
choose among products and fill their shopping trollies. The search
algorithms must be simple and based not only on product names but
also keywords and product categories.
– Simple and effective online payment systems are crucial to the success
of e-commerce. There are several different types of online payment
systems; for example, credit cards (e.g., VISA and MasterCard), e-
wallet (e.g., PayPal), invoice installments (e.g., Klarna), and
cryptocurrencies (e.g., Bitcoin
, Ethereum
, and Ripple
).
E-commerce market can be divided into four categories depending
on whether the buyer or the seller is a professional business (denoted
as “B”) or a private consumer (denoted as “C”). These four types of e-
commerce markets are as follows (see ◘ Fig. 17.5):
– Business-to-Consumer (B2C) e-commerce is the traditional market in
which goods or services are sold online by professional companies to
private consumers. Examples are buying books from Amazon, films
from Netflix, flight tickets from Expedia, computers from Dell, and
broadband subscriptions from a network operator.
– Business-to-Business (B2B) e-commerce is the online trading between
two professional companies. B2C e-commerce and B2B e-commerce
have experienced huge growth during the last decades. The main
difference between B2B and B2C e-commerce is that, in B2C, small
quantities of goods and services are sold to many private consumers,
while in B2B, large quantities of goods and services are sold to a
small number of professional businesses.
– Consumer-to-Consumer (C2C) e-commerce is the online trading
between two private consumers. Examples of C2C e-commerce
companies include eBay, Uber, and Airbnb. In fact, most of the
sharing economies are C2C e-commerce. Advantages of C2C e-
commerce are better utilization of resources and simpler trade
opportunities for second-hand goods.
– Consumer-to-Business (C2B) e-commerce enables private consumers
to sell digital services online to professional companies. This is the
 most recent supplement to e-commerce. One example of C2B e-
commerce is a private consumer’s web page on which manufacturers
and retailers advertise their products. A blogger may have many
viewers on their blog, and manufacturers or providers may find the
blog to be a simple and cheap way to reach a particular audience.

Fig. 17.5 Classification of e-commerce markets. (Authors’ own figure)

Box 17.1 The Sharing Economy

The sharing economy enables consumers to sell access to property,
goods, money, or services to other consumers for a certain fee.
Airbnb and Uber are two of the most well-known companies in the
sharing economy. The sharing economy is an example of C2C e-
commerce since the trade takes place between two consumers.
A more precise term of the sharing economy is access economy
.
This is because, strictly speaking, consumers do not share goods or
services but pay for access to other consumers´ goods and services.
The sharing economy may lead to better utilization of resources since
homes, tools, or cars can be rented out when they are not used by the
owner. An important requirement of sharing economy services is fast
and reliable feedback from the consumers. This is required to build
trust and reputation for those offering access to their assets and
services.
The major difference between the sharing economy and
traditional trade is that the providers are often individuals and not
 companies. In many countries, this means that a different set of laws
and regulations govern the trade. The sharing economy is enabled by
multisided platforms (see ► Chap. 10) and crowdsourcing (see ►
Chap. 7). Sharing economy services use crowdsourcing as a
production model and create value as a value network (see ► Chap.
8) or, more specifically, as a multisided platform. One example is
Uber, in which people (the crowd) offer transportation services to
consumers. Uber does not own taxis or cars but instead mediates
between drivers and passengers. Uber is totally dependent on the
crowd to provide their assets (cars) in its business model.
The sharing economy challenges legal frameworks, especially
labor laws and commercial laws. Uber is, for example, forbidden in
several countries, including Norway, Denmark, and Italy, due to
violation of the laws concerning licensing of professional taxi drivers
(Rhodes, 2017). Another example is Airbnb having met restrictions
in, for example, New York City where private consumers are not
allowed to rent out property on a short-term contract (less than 1
month) when the host is not present. The sharing economy enables
consumers to make profits off assets they own. Sharing services have
been criticized as competing under different terms than established
businesses by circumventing labor protection laws and thereby
providing services with lower costs compared to services produced
by companies using the in-house production model.

17.4 Network Access Markets

Network access is offered jointly by the ISP (commercial) and NP
(technical). This includes access to broadband Internet connections,
Wi-Fi, public mobile networks, telephone services, and messaging
services (e.g., email and SMS/MMS
). These services are integral parts
of the network and do not depend on additional services delivered by
other stakeholders (e.g., ASPs or content providers).
Network access is a fundamental service—also called a
foundational technology—in the digital economy. This is because the
access to and delivery of digital services depends on reliable access to
the Internet. Reliable access to the Internet is supported by a
worldwide ICT infrastructure consisting of optical fibers, wireless base
stations, Internet routers, satellite networks, and other network
resources. Users access the Internet using personal computers, tablets,
set-top boxes, or smartphones. The NP owns and operates the physical
ICT infrastructure supporting the Internet. This includes all kinds of
communication networks and associated management systems and
computing and storage facilities. The ISP buys access to the
infrastructure from the NP and resells this access to consumers and
ASPs.
There are several examples of NPs that are also ISPs. The traditional
incumbent network operator both owns the communication network
and sells telephone services to consumers. To ensure fair competition
among ISPs, national regulation in most countries compels the
incumbent network operator to split the business operations into two
independent parts: one for NP operations and one for ISP operations.
Several national regulators have also forced the incumbent NP to open
the ICT infrastructure for ISPs other than the ISP owned by the
incumbent. These ISPs can lease the ICT infrastructure from the NP on
the same terms as the ISP owned by the incumbent NP.
ISPs that do not own their own network infrastructure are called
virtual network operators (VNOs)—or mobile virtual network operators
(MVNOs) if they offer mobile services. ◘ Figure 17.6 shows an example
of how a VNO (ISPB of Company B) is commercially related to the
network provider (NP) owned by Company A. Both companies A and B
offer services to the consumer (C) through the Internet service
providers ISPA and ISPB, respectively. The major difference between
them is that ISPA is owned by the company that owns the ICT
infrastructure (IP), while ISPB leases access to the same infrastructure
from Company A. See ► Chap. 5 for other aspects concerning VNOs.
Fig. 17.6 Virtual network operator. (Authors’ own figure)

Over-the-top services (OTT

) are media services offered by the ASP
or content provider directly over the network of the NP. OTT services
require access to the Internet delivered by the ISP. However, OTT
services compete with the media and communication services offered
by the ISP itself. This is illustrated in ◘ Fig. 17.7, in which the
architecture of OTT messaging, OTT voice, and OTT film distribution is
presented together with the equivalent services offered by the ISP.

Fig. 17.7 Over-the-top services. (Authors’ own figure)

 A consumer may use a smartphone in combination with Internet
access offered by the ISP and WeChat (an OTT messaging service) as an
alternative to SMS messaging provided by the ISP. A consumer may also
use Internet access combined with voice-over-IP telephone service
offered by an ASP—for example, Skype—as an alternative to the
telephone service offered by a mobile service provider. Similarly, Netflix
may provide the same services directly over the Internet as a provider
of cable television services. The key point of OTT is that the same set of
services that was traditionally offered over a dedicated communication
network can be offered over the Internet without other involvement of
the ISP and the NP than transporting bits.
The major challenge for the ISP is smaller revenue because of
competition from OTT services. This is because the price per bit for the
telephone service has traditionally been orders of magnitude higher
than the price per bit for Internet access. Hence, the ISP will face
reduced revenue as consumers move from, for example, telephone
service to OTT voice services since the price per bit that the ISP charges
the OTT provider for network access is much lower than the price per
bit that the ISP charges the consumer directly. For the consumer, OTT
means significantly lower prices for digital services such as telephony
and messaging. OTT is one step toward the convergence of services (see
► Chap. 3), in which traditional telephone service is replaced by VoIP
and cable television services are replaced by video streaming.
Network access services are close to becoming digital commodities
.
This is because it is almost impossible to differentiate between the
various network access services provided by different network access
suppliers. In commodity markets, price is the most important
differentiator between suppliers. This may also be one reason for the
ongoing price war between suppliers of broadband access and mobile
telephone services.
◘ Figure 17.8 extends ◘ Fig. 17.2 to include device provider
s (DP).
Device providers are providers of user equipment such as
smartphones, laptops, and personal computers. Production and trade of
such devices are not part of the digital economy as defined in this book.
However, they constitute an integral part of the digital service
ecosystem because digital services must be accessed through some sort
of device produced and marketed by device providers. A specific device
is interoperable with both the access services provided by ISPs and the
applications provided by ASPs. This condition may, however, be device
specific. For example, apps available at Apple App Store are only
available for devices provided by Apple. However, technologically
speaking, full interoperability exists between network access, devices,
and applications—any device can be connected to any network and run
any application. This separation of technologies between ISP, device
provider, and ASP is the basic concept of the layered Internet model
described in ► Chap. 4.

Fig. 17.8 Device providers in digital markets. (Authors’ own figure)

17.5 Information Service Markets

Information services are jointly offered by the ASP and the content
provider and are traded in an information service market (Linde &
Stock, 2011). They include content and applications ranging from
simple apps to complex software. Huge amounts of digital content and
applications are available for consumers. ◘ Table 17.1 contains
examples of information services and how they may be categorized.
Table 17.1 Examples of information services. (Authors’ compilation)

Type of service Examples of information service

Social media services Facebook, LinkedIn, Twitter, QZone, VKontakte
Type of service Examples of information service
Music streaming Spotify, Apple Music, Google Play Music, Tidal
Video streaming Netflix, HBO, Amazon Video, YouTube TV
Web browsers Chrome, Safari, Internet Explorer, Firefox
Word-editing software Microsoft Word, Google Docs, Pages
Internet telephony (VoIP) Skype, Google Voice
Messaging WhatsApp, WeChat, Messenger, SMS
Multiplayer online games World of Warcraft,
Starcraft 2
Travel and accommodation TripAdvisor, Citymapper, Uber, Airbnb
Online payment PayPal, Alipay, Google Pay, Apple Pay
Language Google Translate, Duolingo
News NY Times, Google News, Reddit

Some of the services listed in ◘ Table 17.1 are available

internationally, while others have a regional target. Since the marginal
cost of digital services is zero, it is simple to distribute them on the
international marketplace. However, there may be several reasons why
some of them are restricted to certain geographical areas, for example:
– Political regulations (e.g., Facebook is not allowed in China)
– Competition regulations (e.g., Uber is forbidden in several countries)
– Language (e.g., local newspapers)
– Local target (e.g., regional transportation apps)
– Infrastructure limitations (e.g., local network not supporting
broadband access)
– Local information (e.g., road, traffic, and weather conditions)
In an article published in Harvard Business Review in 1998, Josef
Pine and James Gilmore coined the term experience economy (Pine &
Gilmore, 1998)
. Their argument is that people are willing to pay for the
experience of being “engaged” in the product they buy. Several
information services belong to the category of experience goods.
Examples of experience goods are movies, interactive games, music,
and newspaper articles. It is hard to assess the quality of an experience
good in advance since it is difficult for an individual consumer to assess
the quality of a specific music track or a movie before it is purchased. To
give consumers some information about the digital good or service, the
provider may have to give away samples of the product or present
evaluations of the product by professional reviewers or by feedback
from the public. Network access, on the other hand, is classified as a
search good (Nelson, 1970)
. The most important characteristic of a
search good is that the quality of the good can be assessed before it is
purchased. Search goods are more subject to price wars and fierce
competition than experience goods. Search goods are often commodity
markets, while experience goods are markets with monopolistic
competition; see ► Chap. 13.

Definition 17.4 Search Good and Experience Good

Search good is a good with attributes or qualities that can be
evaluated before consumption. Experience good is good with
attributes and qualities that can only be evaluated after
consumption.

When content is produced in-house, as explained in ► Chap. 7, the

content provider producing it is a professional company. One example
is the film industry, which requires huge budgets and a professionalized
mode of operation when making a movie. On the other hand, when
content is produced by crowdsourcing or peer production, the content
provider (CP) and the consumer (C) may be the same entity, called a
prosumer
.

Definition 17.5 Prosumer

Prosumer can be defined as “individuals who consume and produce
value, either for self-consumption or consumption by others, and can
receive implicit or explicit incentives from organizations involved in
the exchange” (Lang et al., 2020). In the digital economy, this
definition may be limited to encompass organizations where
individuals may both produce content for other individuals and
consume content produced by other individuals.

This is illustrated in ◘ Fig. 17.9. There are several business models

available for the ASP:
– The ASP may buy copyrighted content from the prosumer and sell it
back to the prosumer.
– The prosumer may give content for free to the ASP, and the ASP may
offer the content to the prosumer either for free or for a certain fee.
– The ASP may charge the prosumers for using the mediation service,
either as producer, consumer, or both.

Fig. 17.9 Prosumers in digital markets. (Authors’ own figure)

A prosumer may either only produce content, only consume

content, or both. The ASP considers the groups of consumers and
producers as a single stakeholder. YouTube is an example of a company
connecting prosumers.

17.6 Conclusions
Digital markets can be categorized in several ways.
– E-commerce markets include all kinds of trading using the Internet
for product search, purchasing, and payment. The good may be
tangible or digital. In the latter case, the good is also delivered on the
 Internet. One basic requirement for a trade to be categorized as e-
commerce is that the trade is associated by financial transactions.
Examples are Amazon and Spotify.
– Network access markets are the commercial activities associated
with network access and transfer of bits between users. The
stakeholders include owners of infrastructures such as Internet
routers, optical fiber networks, wireless access networks, and
satellite networks. The concept also includes Internet service
providers (ISP) offering the infrastructure services such as Internet
access and 5G mobile services to the customers.
– Information service markets consist of the myriad of services offered
by content providers and application service providers.
Moreover, the information service markets may also be categorized
as:
– Over-the-top services where the provider of the service is bypassing
traditional services offered on the network such as voice, television,
and messaging by hiding them in unspecified Internet packets.
– Search goods where the attributes of the service are known before
consumption.
– Experience goods where the attributes of the service are only known
after consumption.
– Prosumer markets where the users are both producers and
consumers of value.

Questions
1.
Amazon and Alibaba are two of the largest e-commerce  
companies in the world.
(a)
Are Amazon and Alibaba doing B2B, B2C, C2B, or C2C e-  
commerce?
(b)
How are Amazon and Alibaba handling online payments?  
(c)
Have Amazon’s and Alibaba’s business operations  
influenced transaction costs?
2.
Apple Pay is an online payment service.  
(a)
Is Apple Pay an experience good or a search good?  
(b)
Is Apple Pay an OTT service?  
3.
Why can we categorize the users of eBay, Twitter, Airbnb, and  
Wikipedia as prosumers?

Answers
1.
Amazon and Alibaba  
(a)
Amazon is primarily a B2C and Alibaba is primarily a C2C.  
However, Amazon has also enabled C2C on its platforms.
(b)
Alibaba has its own payment system called Alipay. Amazon  
relies on external payment systems.
(c)
Both Amazon and Alibaba have contributed to reduced  
transaction costs in the digital economy. This is because
they allow consumers to search for millions of products on
their websites in a very efficient way compared to
traditional retail businesses, e.g., by using keywords.
2.
Apple Pay  
(a)
Apple Pay is a search good: the attributes of the service are  
known before it is used.
(b)
Apple Pay is not an OTT service but offered in combination  
with other Internet services (e.g., e-commerce).
3. eBay offers a platform on which users may sell goods directly to
other users. Users post messages on Twitter that is read and  
 commented by other users. Airbnb mediates between users

having assets for short-term rental (e.g., vacation homes) and

other users in search for such assets. The users of Wikipedia are
both producers and readers of the information on the
encyclopedia. In all cases, the user may be both producer and
consumer.

References
E-commerce in the United States, Statistics and Facts. Statista. July 7, 2020.

Global Mobile Commerce Forum: Inaugural plenary conference. November 12, 1997.

Lang, B., Dolan, R., Kemper, J., & Northey, G. (2020). Prosumers in time of crisis: Definition,
archetypes and implications. Journal of Service Management, 32(2), 176–189.

Linde, F., & Stock, W. G. (2011). Information markets. A strategic guideline for the I-commerce. De
Gruyter Saur.
[Crossref]

Long, D. (2017, July 5). China’s ecommerce market to pass $1.1tn in 2017. The Drum.

Nelson, P. (1970). Information and consumer behavior. Journal of Political Economy, 78(2), 311.
[Crossref]

Pine, B. J., & Gilmore, J. (1998, July 1). Welcome to the experience economy. Harvard Business
Review.

Popovic, A. (2019, August 21). mCommerce: How mobile commerce is changing the way we do
business. Price2Spy.

Rhodes. (2017, September 22). UBER: Which countries have banned the controversial taxi app.
The Independent.

The impact of Covid-19 on ecommerce. Ecommerce News Europe. January 15, 2021.

Further Reading
Laudon, K. C., & Traver, C. G. (2017). E-commerce 2017: Business, technology, and society.
Pearson.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_18

18. Digital Market Modeling

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Bass model – Innovators – Imitators – Churning – SIR

model – System dynamics

Learning Objectives
After completing this chapter, you should be able to:
– Identify the growth mechanisms of evolving markets.
– Set up departmental mathematical models for simple digital
markets.
– Apply strategic issues such as latency, effects of churning,
growth rate, and inflexion on the evolution of real markets.

18.1 Introduction
This chapter presents quantitative models for the temporal evolution of
digital markets. The chapter requires some basic knowledge of
elementary calculus such as ordinary differential equations and simple
algebraic manipulations. Some of the mathematical derivations are
placed in separate boxes to make the text more easily available also to
those who are less skilled in calculus.
 The objective is to uncover the dynamic behavior of markets that
are common in the digital economy, for example, social media,
interactive games, communication services, and sales of electronic
gadgets. An evolving market is not in an equilibrium state, and standard
supply-demand theories do not apply to these markets. Moreover, in
several of these markets, the marginal cost and the price of products is
zero (e.g., Facebook and Google Search) making supply-demand curves
meaningless.
The purpose of this chapter is to show:
– How the markets for certain products (e.g., durables and certain
digital services) evolve and mature as a function of time (► Sect.
18.2)
– Why competition may, in some cases, lead to winner-takes-all
markets and, in other case, to stable markets shared by several
suppliers (► Sect. 18.3)
– How markets like interactive games grow, mature, and die (► Sect.
18.4)
The temporal evolution of the market can, to a first approximation,
be modeled using single first-order differential equations or coupled
sets of such equations. For simplicity, all markets that are considered
consist of a fixed number, N, of potential customers buying the good;
that is, market variations owing to births and deaths processes are
ignored. The equations then become simpler, and the solutions are
easier to understand. The simplification does not alter the validity and
generality of the conclusions.
In some markets, eventually all potential customers have purchased
the good at some time, and no more sales take place. It is also assumed
that there are no other saturation effects (e.g., insufficient supply)
influencing the likelihood that a product is purchased. There are several
examples of services that have evolved in this way—for example,
mobile phone subscriptions and Internet access. In both cases, no
significant saturation effects caused by overload in the technical
infrastructure have been observed during the evolution of these
networks. Similar observations are made regarding the evolution of
several social media services—the providers of the services seem to be
able to put up enough capacity to avoid the saturation effects that
moderate the evolution of the service.
We start with analyzing markets using the market diffusion
equation developed by Frank Bass in 1969 (Bass, 1969). There are
several reasons for this:
– The Bass diffusion model describes rather well the evolution of
markets for durable commodity products where every household
usually needs one item of each, for example, refrigerators, home
freezers, stoves, and lawn mowers. The model is also valid for social
media such as Facebook and Twitter and for subscription-based
digital services such as Internet access, Spotify, and Netflix.
– The Bass diffusion model allows simple analytic solutions from
which we may draw important conclusions concerning the temporal
evolution of the market.
– The model can also be used in more complex cases such as
competition and online games to describe processes such as
customer churn and the rate by which customers are leaving the
service.

18.2 Bass Diffusion Model

◘ Figure 18.1 shows a simple market model for durables (e.g.,
refrigerators and radios) or goods where only one unit of the good is
needed (e.g., newspaper subscription). The model is called the Bass
diffusion model. Examples of applications of the Bass diffusion model in
the digital domain include smartphone subscriptions, Facebook
accounts, and Twitter accounts. The container (or compartment) to the
left in the figure represents individuals who have not yet purchased the
good (potential customers). There are N − B such individuals, in which
N is the total population. The container (or compartment) to the right
represents individuals who have already purchased the good. There are
B such customers. Note that the total population (potential adopters
plus adopters) is assumed to be constant and equal to N and that
anyone having bought the good keeps it indefinitely. This implies that
birth or death processes are not included in the model. To be able to
solve the equations using analytic tools, it is also assumed that all flow
parameters are constants.
Fig. 18.1 The Bass diffusion model. (Authors’ own figure)

When individuals buy a good, they become customers and move

from the potential customers container to the customers container. This
is the basic dynamics of the model.
There are two flows of new buyers of the good:
– Flow of innovators
consists of individuals who buy the good
independently of who else has bought it. These are the spontaneous
buyers—also referred to as early adopters. The strength of the flow is
p(N − B) in which the intensity by which an arbitrary individual will
buy the good is p. The parameter p is called the coefficient of
innovation
.
– Flow of imitators
consists of individuals who buy the good because
others have bought it. These individuals are also called late followers
(or just followers) or stimulated buyers. The strength of the flow is
qB(N − B) in which qB is the intensity by which an individual is
stimulated into buying the good, expressing that the intensity by
which the imitator will buy the good is proportional to the current
number of customers. The parameter q is called the coefficient of
imitation
.
 The model also contains a feedback loop representing the network
effect. The strength of the feedback is proportional to the number of
customers who have already bought the good. This is a network effect
of the Metcalfe type; see ► Chap. 9. ◘ Table 18.1 summarizes the
parameters used in the Bass diffusion model.
Table 18.1 Parameters in the Bass diffusion model. (Authors’ compilation)

Parameter Description
N Total population
B Individuals that have bought the product or are using the service at time t (i.e.,
the current customers)
p Coefficient of innovation
q Coefficient of imitation
t Time

The total flow of customers at any instant of time, dB/dt, is equal to

the sum of innovators and imitators buying the service at that instant of
time. The differential equation for the market dynamics follows then
directly from ◘ Fig. 18.1:

This is the Bass diffusion equation. Note that the term p + qB is the
intensity by which an item is sold in the infinitesimal period dt. dB/dt
is, therefore, the demand for the good at time t since the demand is, by
definition, the same as items sold per unit of time. The Bass equation is
solved in ► Box 18.1.
◘ Figure 18.2 shows the solution of the Bass equation for a total
population of N = 106 individuals, B0 = 104 initial customers, the
coefficient of innovation p = 0.03, and the coefficient of imitation q = 3.8
× 10−7. The values of p and q are based on the typical and average
values found in the paper by Mahajan et al. (1995).
Fig. 18.2 Plot of the Bass diffusion model. (Authors’ own figure)

There are two special cases of the Bass equation:

– If p = 0, there are only imitators, and the solution of the Bass
equation is reduced to the logistic distribution (see, e.g., the
Wikipedia article for the definition of the logistic distribution):

– If q = 0, there are only innovators, and the solution of the Bass

equation is reduced to the exponential distribution:
 Note that if there are no innovators (early adopters) but only
imitators (i.e., p = 0), the solution of the differential equation is B = 0 for
all t if B0 = 0; that is, no one will ever buy the product. Therefore, B0 > 0
for a non-zero solution to exist if p = 0; that is, a customer base must
exist before the sales begin. This requirement is not necessary for the
case of only innovators. In this case, customers will buy the good even if
there are no initial customers.
◘ Figure 18.3 shows the solution of the Bass equation for the two
cases with only imitators (p = 0, q = 3.8 × 10−7) and only innovators (p =
0.05, q = 0). For both graphs, the total population is N = 106, and the
initial number of customers is B0 = 104.

Fig. 18.3 The Bass diffusion model with only imitators and only innovators. (Authors’ own
figure)
 The condition that the Bass equation produces an S-curve is that
there is an inflexion point
somewhere on the curve. An inflexion point is
a point in which the tangent to the curve has a maximum (or minimum)
value; that is, the increase in the growth rate changes from positive to
negative (or vice versa). At this point, the second derivative of the curve
is zero. How the inflexion point is found is shown in ► Box 18.2.
One of the most important issues when introducing a new product
is the time it takes until enough customers have purchased the good so
that the business has become profitable. This may be called the latency
period for market penetration. It is reasonable to define the latency
period to be the time it takes to reach 10% of the full market size (T10).
For a market with only imitators (p = 0 in the Bass equation), we find
by simple algebra that:

in which T50 is the time it takes to reach 50% of the potential market. ◘
Table 18.2 shows the latency period for several values of B0 (the
number of customers that must be captured before the product is
launched). In the table, the time to reach 50% of the market, T50, is 5
years.
Table 18.2 Latency periods vs initial customer base. (Authors’ compilation)

B0/N T10/T50 T10 for T50 = 5 years

0.001 0.67 3 years and 4 months

0.005 0.58 2 years and 11 months
0.01 0.52 2 years and 7 months
0.02 0.44 2 years and 2 months
0.04 0.31 1 year and 6 months

If T50 = 5 years and B0/N = 0.001, the latency period is 3 years and 4
months. If B0/N = 0.01, the latency period is still 2 years and 7 months.
In markets without (or with very few) spontaneous buyers, the latency
period is long, and the supplier may choose to terminate the service
before the network effects become significant. This is the strategic
dilemma in markets with strong network effects and, in which, there is
a minor incentive for users to spontaneously join the service. The
service may then be terminated before the market has started to
mature. See also ► Chap. 9 where the problem of long latency period is
discussed in the context of network effects.
If there are only innovators (i.e., q = 0) and B0 = 0, the solution
reduces to:

In this case the latency period (T10) is given by:

For T50 = 5 years, the latency period is only 9 months. In the case
with only innovators, the market increases linearly, B(t) ≈ pNt, for small
t. On the other hand, if there are only imitators, the market increases
exponentially for small t, B ≈ B0eqNt. The importance of this observation
is that initially (i.e., for small t) exponential growth is much slower than
linear growth. This is the origin of long latency period in markets with
only imitators.
On the other hand, observe that the time for the market to increase
from 50% to 60% is only 6 months in the case of only imitators and
B0/N = 0.01, but more than four times as long without feedback (i.e.,
only innovators). The conclusions of this discussion are shown in ◘ Fig.
18.4 and summarized as follows:
– If all customers are innovators, then the latency period is short;
however, the time to capture market shares above 50% is long.
– If all customers are imitators, the latency period is long; however, the
time to capture market shares above 50% is short.
Fig. 18.4 Latency period in the Bass diffusion model. (Authors’ own figure)

There are two strategic dilemmas in a market with only imitators:

– As previously discussed, the market will not start growing unless
there are some initial customers. The problem for the supplier is,
then, to establish an initial pool of customers so that the growth
process will start.
– Even with an initial pool of customers, the growth rate may initially
be so slow that the supplier will terminate the service before it takes
off.
Facebook is an example of a service in which there are very few
innovators, since the reason to use the service is to interact socially
with other users; that is, p ≈ 0 for Facebook. The service was launched
at the campus of Harvard University, building up a small initial user
group among students. It took about 5 years (from 2003 to 2008)
before the market share really started to increase and Facebook started
to become a dominating social networking service (Roberts, 2017).
Other social networking services, such as LinkedIn, also grew slowly
initially.

Box 18.1 Solution of the Bass Equation

The Bass equation is separable since dB = (p + qB)(N − B)dt, resulting
in the following equation where the dependent variable B and the
time variable t have been moved to each side of the equation:

Expanding the left-hand side of the equation and multiplying both

sides by p + qN results in:

Integrating both sides of this equation term by term gives:

By simple algebraic manipulations we find:

in which c is the constant of integration. Observe that for t = 0, B(0) =

B0 (the initial condition), so by setting t = 0 and B = B0 in the
equation, the constant of integration is easily found:

Inserting this and solving for B finally results in the solution of

the Bass equation:
 in which B0 = B(0) is the initial number of individuals possessing the
good. These may be individuals who have attained the good as part of
a marketing promotion, a product test, or together with a
complementary product (e.g., the SMS attached to mobile phones).
The significance of these customers is discussed in the main text.

Box 18.2 Finding the Inflexion Point

The inflexion point is the point where the second derivative of the
solution of the Bass equation vanishes, i.e., d2B/dt2 = 0. The second
derivative of the solution to the Bass equation is:

The condition for the existence of an inflexion point is:

Since dB/dt > 0, this leads to the linear equation q(N − B) − (p +

qB) = 0 with solution:

in which Binfl is the value of B at the inflexion point. Observe that

there is an inflexion point on the curve for B ≥ 0 provided that p < qN.
If this condition is not fulfilled, there is no inflexion point on the
positive part of the curve.

18.3 Model for Markets with Competition and

Churning
◘ Figure 18.5 shows a model for the competition between two
suppliers—Supplier 1 and Supplier 2—offering the same service; for
example, mobile communications. The model consists of three
customer states: potential customers, customers of Supplier 1, and
customers of Supplier 2. Furthermore, the model consists of four flows:
– The flow of new customers from potential customers to Supplier 1.
– The flow of new customers from potential customers to Supplier 2.
– The flow of customers from Supplier 1 to Supplier 2.
– The flow of customers from Supplier 2 to Supplier 1.

Fig. 18.5 Model of two competing suppliers with churning. (Authors’ own figure)

The latter two flows are called churning

. The model also includes
four feedback loops as shown in the figure.
The rate of new customers choosing Supplier 1 is (p1 + q1B1)(N − B1
− B2), and the rate of new customers choosing Supplier 2 is (p2 + q2B2)
(N − B1 − B2). Here, the Bass equation is used to express the dynamics
of the flows.
The net number of customers churning from Supplier 2 to Supplier
1 is C21 = B2(c2 + d2B1) − B1(c1 + d1B2). The first term is the rate at
which Supplier 1 gains customers from Supplier 2, and the second term
is the rate at which Supplier 1 loses customers to Supplier 2. The net
number of customers churning to Supplier 2 is C12 = − C21 = B1(c1 +
d1B2) − B2(c2 + d2B1) since the churning process does not create new
customers (the net result of the two churning flows must be zero). The
parameters c1 and c2 are denoted coefficients of spontaneous churning,
and the parameters d1 and d2 are denoted coefficients of stimulated
churning (or imitated churning). Again, the Bass equation is applied to
the churning flows. Imitated churning constitutes churners that switch
suppliers because of the size of the competing supplier.
The dotted lines in the figure are the feedback loops representing
the network effect that stimulates imitators into choosing a supplier or
stimulates customers into churning to the other supplier.
Putting all this together results in the following coupled set of
dynamic market equations:

There is little hope to solve these nonlinear differential equations

analytically, except in a few special cases. However, we may still draw
some important conclusions concerning the long-term evolution of the
market without solving the set of differential equations as explained in
► Box 18.3.
In the case of two suppliers, there are two special cases that can be
observed in different markets:
– If there is only stimulated churning (c1 = c2 = 0) and d1 and d2 are
independent of time, then the final state is B1 = 0, B2 = N if d1 > d2, or
B1 = N, B2 = 0 if d1 < d2. These are, then, winner-takes-all markets
leading to de facto monopolies (e.g., Facebook vs Myspace or VHS vs
Betamax).
– If there is only spontaneous churning (d1 = d2 = 0), then it follows
most easily directly from the churning conditions (or from the above
equations by letting (d1 − d2) → 0) that the market ends up in the
stable state with the following steady-state distribution of
customers:
 This case may apply to mobile communications where competitors
have rather stable market shares over long periods of time. We see that
this state depends only on the churning parameters and is independent
on how the market grows before it is saturated. In this simple model,
the spontaneous churning coefficients are treated as constants.
However, in actual markets, they may be complex time-dependent
functions of prices, service content, user experience, user preferences,
and so on. The market shares will then become fluctuating functions of
time which, in some cases, may lead to winner-takes-all markets, for
example, if churning only takes place from one competitor to the other
(e.g., if c1 = 0, c2 > 0, then B1 = N and B2 = 0).
It is easy to extend the model to more than two competitors. If there
is only spontaneous churning, it is feasible to find analytic expressions
for the stable end state of the market for any number of competitors,
though it is numerically cumbersome to calculate the exact values if
there are more than three competitors. On the other hand, if there is
only stimulated churning (and no spontaneous churning), then the
market will eventually end up in a state in which one of the competitors
has captured the whole market. This is also a winner-takes-all market.
Note that in these models, the assumption is that the average
churning probability is constant. In real systems, this is not the case,
and it is reasonable to assume that the churning probability is a
complex, fluctuating function of time depending on parameters such as
price, loyalty, technical quality, customer laziness, or other mechanisms
which may motivate the user to churn or not to churn to another
supplier. The motivation of this chapter is not to describe why users
may churn but to show that churning may result in a number of final
market states ranging from de facto monopolies to rather stable
markets shared by two or more suppliers. The theory also shows that
the long-term evolution is path dependent, where the path the market
evolution will follow depends on all the parameters just mentioned (see
also ► Chap. 11).
 Box 18.3 Market Stability and Churning
In the long run, all potential customers have become customers of
either Supplier 1 or Supplier 2, and there are no more potential
customers left. This is, for example, the case in the mobile phone
market in several countries (this has nothing to do with the sales of
mobile phones but with the total number of mobile subscriptions).
This means that B1 + B2 = N. A steady state solution implies,
moreover, that dB1/dt = dB2/dt = 0; that is, there is no net flow of
customers in the steady state. In the steady state, there is, therefore,
no net churning (i.e., C12 = − C21 = 0)
, which results in the solution of
the quadratic equation B1(c1 + d1B2) = B2(c2 + d2B1), in which B1 + B2
= N, for the final state of the market. This means that a potential
customer has either become a subscriber of Supplier 1 or Supplier 2
and that the churning rates of the two suppliers are equal. The
general solution is then:

From these observations, we draw some important conclusions in

the main text.

18.4 Models for Massive Multiplayer Online

Games
◘ Figure 18.6 is a simple model for a massively multiplayer online game
(MMOG
) such as World of Warcraft
. The model may also be used to
analyze services in which the users may leave the service with a certain
probability. Examples include social networking services and
newspaper subscriptions. An individual or a player may be in one of
three possible states: potential player (B), player (P), or a player who
has quit the game (Quitter) (Q).
Fig. 18.6 Model of a massively multiplayer online game. (Authors’ own figure)

There are three flows, in which it is assumed that the rate of each
flow obeys the Bass equation:
– New players enter the game with rate (p + qP)B.
– Players leave the game with rate (r + sQ)P.
– Players rejoin the game with rate (u + vP)Q.
The dotted lines in the figure show the network effect. For
simplicity, we will call this model the BPQ model (Øverby & Audestad,
2019).
The coupled set of differential equations is now:

Adding the three equations results in dB/dt + dP/dt + dQ/dt = 0.

This leads to the obvious conservation law B + P + Q = N, in which N is
the total population of potential players. As usual, the model is
simplified by assuming that N is constant (no birth or death processes).
The number of independent differential equations is then reduced to
two since the conservation equation can be used to eliminate one of
them. These equations can then easily be transformed into a single,
rather intractable, nonlinear second-order differential equation for P.
There are a few cases in which analytic solutions can be found.
However, we shall not pursue this here.
◘ Figure 18.7 is an example of a typical solution of the differential
equations of the BPQ model. The differential equations were solved
using numerical methods. The figure shows the share of the population
that are potential players (B), players (P), or quitters (Q) as a function of
time. Observe that the sum of these categories of individuals is always
100%. At t = 0, there are only potential players and no players or
quitters (i.e., B = 100% and P = Q = 0%). As time increases, more and
more potential players become players, who, after some time, leave the
game and become quitters. However, quitters may rejoin the game and
become players again. Eventually (in the figure, for t > 90), all
individuals are either players or quitters.
Fig. 18.7 Plot of the BPQ model. (Authors’ own figure)

From the form of the differential equations, important conclusions

can be drawn without solving the equations:
– If there are no innovators among the potential players, the time it
takes for the game to reach sufficient popularity may be long in the
same way as in the Bass model with only imitators. The game may
then be prematurely withdrawn from the market.
– To prolong the lifetime of the game, quitters must be stimulated to
rejoin the game. This requires frequently updating of the game with
 interesting new features.
– Updating of the game may also stimulate current players to continue
to play, thereby reducing the rate by which payers are leaving the
game.

Box 18.4 The SIR Model

The simplest model of interactive games is similar to the model used
in biological sciences to describe how epidemic diseases spread in a
population, the SIR model (Murray, 2002).
The SIR model consists of three groups of individuals: Susceptible
(S), Infected (I) and Recovered (R). This corresponds to B, P, and Q,
respectively, in the BPQ model. The flow of new infected individuals
in the SIR model depends only on network effects—someone must
infect you; the flow of recovered individuals does not depend on
network effects—you recover independently of how anybody else
recovers. The parameter β−1 denotes the time between contacts
which is required for the transmission of the disease. The parameter
γ−1 is the time it takes to recover from the disease. ◘ Figure 18.8
shows the SIR model. The resulting differential equations for the SIR
model are:
Fig. 18.8 The SIR model. (Authors’ own figure)

The set of differential equations is nonlinear and does not have a

closed-form solution; however, the solution is easily found by
numerical integration. The most important conclusion is that,
initially, the number of infected increases very slowly (as in the Bass
model with only imitators) and then to increase very rapidly.
The spread of the COVID-19 pandemic follows the simple SIR
model. Countries have implemented several countermeasures to
reduce the spread of the disease. The differential equation shows
that this is achieved by reducing the term βSI. Examples of
countermeasures that reduce this product are:
– Increased social distance and hand washing reduces β.
– Isolation of particularly vulnerable people, curfews, and
prohibiting many people to assemble in places where social
distance cannot be upheld reduces S.
– Isolating infected and possibly infected people reduces I.
The SIR model was published by A. G. McKendrick and W. O.
Kermack in a series of papers in the period from 1927 to 1933. The
SIR model is the basis for more advanced compartmental models in
epidemiology, such as the SIS model, MSIR model, and the SEIR
model. The major differences between these models are the number
of compartments (user groups) and the interaction between them.
Compartmental models have inspired academics to develop
similar models for the evolution of digital goods and services in the
digital economy—the Bass model, the model with competition and
churning, and the BPQ model are examples of such compartmental
models in the digital economy. J. Cannarella and J. A. Spechler
modified the SIR model by adding positive feedback to the recovery
mechanism. They called the model the irSIR model, where “ir” stands
for “infectious recovery,” and used it to forecast the evolution of
Facebook and Myspace (Cannarella & Spechler, 2014)
. The model
fitted well with the evolution of Myspace but not for Facebook. The
growth rate of Facebook had started to decline during 2013, and
from this data, they predicted (in 2014) a rapid decline in Facebook
users starting from 2017. The growth rate of Facebook regained its
original speed in 2015, so the prediction was based on too little data.
 This is similar to the problem the health authorities have to predict
the evolution of COVID-19.

18.5 Analysis of Real Markets

As shown, the temporal evolution of markets can be described
mathematically using coupled sets of ordinary differential equations. In
most cases, these equations cannot be solved using analytical methods.
In the few cases in which analytical solutions can be found, it is
necessary to treat all flow parameters as constants. Even the simple
Bass equation cannot be solved analytically if the coefficients of
innovation and imitation are functions of time.
Real markets are much more complicated than the three simple
cases described in this chapter: system parameters are not constants
but may be complex functions of time and the number of current
customers; the model itself may be more complex containing additional
customer states (or compartments) and flows; and there may be more
complex feedbacks from the market regulating the flows between
states. Such complex cases can be analyzed using system dynamics. The
method is briefly outlined in ► Box 18.5. The interested reader is
advised to consult the specialized literature on system dynamics (e.g.,
the two books listed below) to learn more about this important method
for the analysis of complex systems.

Box 18.5 System Dynamic Models

System dynamics is based on essentially the same method as
differential equations. However, instead of solving the equations
using standard analytical or numerical methods, the equations are
converted into a dynamic simulation model. The strength of system
dynamic simulations is that the complete simulation model can be
compiled into executable software programs directly from the
graphical description of the model. There are several commercially
available software packages for system dynamic modeling, all of
them based on graphical description of the system. The task for the
designer is then to develop the graphical model.
The method was developed by Jay Forester during the late 1950s
(Forester, 1971). The first major application of the method was the
project at MIT resulting in the book The Limits to Growth published in
1972 (Meadows et al., 1972).
The system dynamic model allows us to treat all system
parameters as continuous or discrete functions of time and simulate
cases that are far out-of-reach using differential equations. ◘ Figure
18.9 shows a system dynamic model of the Bass equation,
demonstrating that the simulation model is identical to the
differential equation in ► Sect. 18.1
.

Fig. 18.9 System dynamic model for the Bass equation. (Authors’ own figure)

In system dynamics, the aggregates of people, things, or money

are called stocks. In the Bass model, there are two stocks, potential
adopters and adopters, and there is one flow from potential adopters
to adopters. There are three functions:
– Innovators having a flow rate of p(N − B). The function is realized
by the multiplication operation × with inputs p and N − B.
– Imitators having a flow rate of qB(N − B). The input to the
multiplication function × is in this case q, N − B, and B.
– New adopters which is the sum of innovators and imitators; that is,
the flow rate of new adopters is (p + qB)(N − B).
 Setting adopters equal to B, the flow of new adopters per unit of
time equal to and potential adopters equal to N − B, the Bass
equation of ► Sect. 18.1 is deduced. When the simulation starts, the
stock of adopters may be empty or contain an initial number of
adopters, B0. The initial stock of potential adopters is, then, either N
or N − B0.

18.6 Conclusions
This chapter shows how some important observations in the digital
economy can be substantiated using simple mathematical tools.
The solution of the Bass equation pinpoints some strategic
dilemmas:
– If all customers are innovators, then the latency period is short;
however, the time to capture market shares above 50% is long. This
implies that the total market is smaller than anticipated.
– If all customers are imitators, the latency period is long; however, the
time to capture market shares above 50% is short. The product may
then be prematurely withdrawn from the market.
The differential equations for markets with competitors show that if
there is stimulated customer churning between the competitors, the
market ends up as a de facto monopoly. Which competitor captures the
whole market is arbitrary depending on events during the evolution of
the market. The equation also shows that there are cases where the
market is shared between several competitors. In these cases, the
churning is spontaneous and independent of the market shares of the
competitors.
The form of the differential equations for interactive online games
shows that strategy to extend the lifetime of the game is to reduce the
rate by which players leaves the game and increasing the rate by which
earlier players rejoins the game. The strategy is then to identify how
these rates can be manipulated, for example, by adding new features
and improving other aspects of the game.
Finally, more complex markets can be analyzed using system
dynamics. In these models, the market parameters can be constants,
depend on time, be discrete, and so on.
Questions
1.
Consider the Bass equation and the two graphs in ◘ Fig. 18.3 (B0  
= 104, N = 106, p = 0.05, and qN = 0.38). Assume that the graphs
represent customer adoption to a social media service.
(a)
How large is the initial market share (i.e., at time t = 0)?  
(b)
Consider the case with only imitators (p = 0). What is this  
distribution called?
(c)
How can the Bass equation be expressed when p = 0?  
(d)
How many customers have adopted the service at the  
inflexion point when p = 0
? And when p = 0.05?
(e)
At what time does the inflexion point occur when p = 0?  
(f)
What is the rate of new customers joining the service at the  
inflexion point when p = 0?
2.
Consider the Bass equation when there are only innovators (q =  
0).
(a)
What is the solution of the Bass equation for q = 0 and B0 >  
0?
(b)
Find the inflexion point in this case (if any).  
(c)
Show that the market increases linearly for small t.  
3.
Consider two digital services satisfying the model for markets  
with competition and churning. Assume that c1 = c2 ≠ 0 and d2 >
d1. Which of the following best describes the market state in the
long run: B1 = B2, B1 > B2, or B1 < B2? Explain why.
Answers
1.
Bass equation.  
(a)
At time t = 0, the initial market share is B0/N = 104/106 =  
1%.
(b)
The distribution is called the logistic distribution.  
(c)
When p = 0, the solution of the Bass equation is:  

(d)
The number of customers at the inflexion point for p = 0
is  
Binfl = N/2 = 500,000. When p = 0.05, the number is Binfl =
(qN − p)/2q = 434.211.
(e)
Solving the Bass equation with p = 0 for t gives:  

Inserting the value for Binfl and the other parameters,

we find that the inflexion occurs at time t = 12.09.
(f)
The rate of new customers is dB/dt = qB(N − B). When B =  
Binfl = 500,000, this gives dB/dt = 3, 8 × 10−7 × 500,0002 =
95,000 customers/year.
2. Bass equation for q = 0.
(a)
 
For q = 0, the solution of the Bass equation is B(t) = N − (N  
− Bo)e−pt.
(b) We easily find that d2B/dt2 = − p2(N − B0)e−pt < 0 for all t.
Therefore there is no inflexion point in this case
 
 Therefore, there is no inflexion point in this case.

(c)
 
Expanding the equation for B(t) as a series, and only
keeping terms to the first order in t gives:

This is the equation for a straight line.

3.
The correct answer is B1 > B2.  

References
Bass, F. M. (1969). A new product growth model for consumer durables. Management Science,
15(5), 215–227.

Cannarella, J., & Spechler, J. A. (2014) Epidemiological modeling of online social network
dynamics. ArXiv 1401.4208.

Forester, J. (1971). Counterintuitive behavior of social systems. Theory and Decisions, 2, 109–
140.

Mahajan, V., Muller, E., & Bass, F. (1995). Diffusion of new products: Empirical generalizations
and managerial uses. Marketing Science, 14(3), G79–G88.

Meadows, D. H., Meadows, D. L., Randers, J., & Behrens, W. (1972). The limits to growth. Potomac
Associates – Universe Books.

Murray, J. D. (2002). Mathematical biology. I: An introduction (3rd ed.). Springer.

[Crossref]

Øverby, H., & Audestad, J. A. (2019). Temporal market evolution of interactive games. SSRN.

Roberts, P. (2017). The most important Facebook statistics for 2017. Our Social Times,
Cambridge.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_19

19. Digital Business Models

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords BMC – SRM

Learning Objectives
After completing this chapter, you should be able to:
– Use the business model canvas to model digital businesses.
– Use the stakeholder relationship model to model network
effects in digital industries.
– Explain the business model of some popular digital services,
including World of Warcraft, Spotify, Facebook, Wikipedia, and
Airbnb.

19.1 Modeling Concepts

Business modeling as a tool for strategic business analysis was
developed in the beginning of the twentieth century. One of the first
business models was the “bait-and-hook” business model, in which a
basic product is first offered for free or for a very low price, while
necessary add-ons are sold for a high price. One example is the razor
(bait) and blades (hook). Another example from the digital economy is
the Adobe PDF software suite, in which the software needed to read
PDFs is free (bait), while the software needed to generate a PDF is
expensive (hook).
Leading developers of new business models in the twentieth
century include companies such as Ford, Toyota, Wal-Mart, FedEx, and
Dell. Dell revolutionized the way computers are sold to customers by
bypassing retailers. Customers place the order directly with Dell who
assembles the computer and ships it directly to the customer (just-in-
time production). This business model was enabled and fueled by the
mass adoption of the Internet. A common feature of these new business
models is that they often exploited breakthroughs in novel
technologies.
A business model describes how an organization (or company)
creates, delivers, captures, and keeps value (Ovens, 2015)
. The model
consists of items such as:
– How the organization earns money and controls expenditures
– Economic and organizational aspects of the organization
– Competition and market evolution
– Interrelationships between the organization and its customers, key
partners, and other stakeholders
Business models are used as a strategic tool to identify key aspects
of an organization’s business operations, that is, the most important
logics and operations that the organization relies on to earn money. The
main purpose of a business model is to describe the current state of the
organization. However, business models are also used as input to
identify key strategic actions required for competitive evolution of the
organization. The business model may be detailed and contain every
aspect of the organization’s business operation. It may also be used to
analyze and describe only parts of an organization’s business
operations. Business models can be applied to all kinds of
organizations.

Definition 19.1 Business Model

A business model describes how an organization creates, delivers,
captures, and keeps value.
In the definition of business model, the concept of value appears. Value,
in this context, is an abstract concept comprising more than just the
monetary value of a product. Value might, for instance, comprise new
features enabled by novel technologies, improved performance,
exclusive design, and usability of the product. The core of a business
model is the value that the organization creates. An organization can
justify its existence by creating value, delivering value to customers,
and supporting efficient mechanisms for gaining revenue from the
value created. In addition to this, the organization must also defend this
value from competing organizations. This may be called the business
process as defined in ► Definition 19.2.

Definition 19.2 Business Process

The business process consists of the following elements:
– The organization creates value by solving problems or satisfying
customer needs. A manufacturer of mobile phones may offer a
low-cost mobile phone to people with slender means and
expensive mobile phones with exclusive design and additional
features to techno-freaks. Built on the same basic technology, the
manufacturer may then satisfy several user segments at the same
time.
– The organization delivers value to the customer through either
physical or digital channels. Value delivery in this context refers to
how the organization transfers the value created to the customer.
An example of value delivery is to ship a purchased mobile phone
to the customer via postal services. Another example is to update
the phone by downloading new software.
– The organization captures value when its customers pay for the
good or reward the organization by other means. The revenue
covers costs and creates profit for the organization. An example of
captured value is payment for the mobile phone and fees for using
particular features such as apps.
– The organization keeps value by, for example, improving the
design and functionality of the product and by mechanisms
protecting against competition. Protection mechanisms are lock-
in of customers to products or services that are difficult to copy,
 protecting the product by copyrights and patents, or selling it for
prices that cannot be matched by competitors.

A digital business model is a business model applied to digital goods or

services. Digital business models are facilitated by the widespread use
of the Internet, mobile technologies, smartphones, and fast and small
computers. As digital technologies mature and become adopted by the
population, novel business models that build on these technologies are
created; for example, a completely new market for app design emerged
after the 4G mobile technology was introduced in 2010. This may lead
to market disruption where new business models may replace existing
ones or create entirely new markets (Christensen, 1997). One example
is the app market just mentioned. Another example is the music
industry which has been radically changed after the introduction of
small and powerful music devices (e.g., the iPod) and online streaming
services (e.g., Spotify).

Example 19.1 Funcom’s Business Model

Funcom is a Norwegian company developing video games, known for
titles such as Anarchy Online
(2001), Age of Conan
(2008), and The
Secret World
(2012). At its peak in 2008, Funcom had over 680
employees in several countries. The company focused on developing
big games with large budgets, thereby competing with successful and
market-leading games such as Blizzard’s World of Warcraft
.
The year 2012 was critical for Funcom. The company launched the
game The Secret World
with little financial success. Funcom had to cut
its staff by several hundred people. The Secret World did not attract as
many players as expected. One of the most important reasons for this
was the choice of business model. The Secret World was initially using
a subscription-based business model. This was, in part, inspired by
the success of the business model of Blizzard’s World of Warcraft
.
Blizzard had already used the subscription-based business model
successfully from the launch of World of Warcraft in 2004. However, in
2012, many computer gamers switched to a free-based business
model used on most games available on social media networks such
as Facebook. Users did not accept to pay a monthly fee for playing a
game anymore. In the end, Funcom had to change the business model
 on The Secret World from subscription-based to buy-to-play to align
with market and user trends.

Recent years have witnessed an increased complexity of business

model designs. This is primarily due to globalization and widespread
competition, more complex and refined technologies, extensive use of
ICT, and more complex organizations. A digital company need not own
the goods they are selling as exemplified by the business models of
eBay, Airbnb, and Uber. The business model of these companies is to
offer a platform for mediation between buyers and sellers of tangible or
intangible goods—eBay does not own any of the physical goods they
are selling even though millions of items are sold through eBay every
year. Airbnb is the world’s largest hospitality service without owning
any hotels or property. Uber provides ridesharing services without
owning any cars.
The business of connecting customers or different user groups is
perhaps the most important contribution of ICT in business modeling.
This has similarities to the transition from business models for value
chains to business models for value networks. While the best way to
model companies in the industrial economy is the value chain, the best
way to model companies in the digital economy is the value network
(see ► Chap. 8). The business model of a company may also need
revision and updating over time and must sometimes be completely
rewritten due to changes in customer behavior, social trends, economic
boundary conditions, and technological evolution. Failure to update
business models may result in reduced profits and, eventually,
bankruptcy.

19.2 The Business Model Canvas

Alexander Osterwalder proposed the business model ontology as part
of his Ph.D. thesis in 2005. Later, this business model ontology was
refined to become the business model canvas (BMC) (Osterwalder &
Pigneur, 2010). The BMC is a framework for describing business
models. It can be applied to all kind of businesses, including digital
businesses. The BMC is based on describing nine central building
blocks of a business and modeling the relationships between these
building blocks. ◘ Figure 19.1 shows an outline of the BMC and how the
BMC is visualized in this book. The different building blocks in the BMC
are described in ◘ Table 19.1. The BMC has been applied to several
digital businesses to uncover business operations and relationships
between stakeholders. In addition to the BMC, there are several other
frameworks for describing businesses. These frameworks focus on
slightly different aspects of the business operations of a company;
however, they mostly agree on the core concepts of the business model.

Fig. 19.1 The business model canvas. (Authors’ own figure)

Table 19.1 The nine building blocks of the BMC. (Authors’ compilation)

Building Description
block
Value Describes the values and benefits created by the organization that are offered
proposition to one or more customer segments. The value proposition may comprise one or
several different propositions targeting different customer segments. The value
proposition describes the goods and services that the organization produces
and delivers to customers, as well as the benefits that customers get by buying
them
Building Description
block
Customer Identifies one or several customer segments targeted by the organization’s
segments value proposition. Different customer segments may have different roles in the
BMC
Channels Describes how the value proposition is transferred to the customer segments
Customer Defines the organization’s relationship with the customer segments
relationships
Key Describes the key activities needed to create and offer the value proposition
activities
Key Classifies the key resources needed to support the key activities and to create
resources and offer the value proposition to the customer segments
Key partners Identifies the key partners required for creating and offering the value
proposition to the customer segments
Cost Identifies the elements that contribute to the cost of the organization, including
structure the cost of the various key activities
Revenue Describes how the different customer segments contribute to the
streams organization’s income

The nine building blocks of the BMC can be divided into three
groups: value proposition, value turnover, and value generation.
Value proposition is the core building block of the BMC. The
advantages of well-defined value propositions are:
– Upholding a clear focus on the fundamental activities of the business
– Identifying and maintaining any core competencies that the company
may possess
– Precise targeting of the production toward products that the users
will have and, thereby, avoiding production of goods that nobody will
buy
– Willingness to change direction as processes, technologies, and
products are substituted by new ones
– Willingness to change direction based on feedback from the users as
the market evolves
– Effective marketing that focuses on user needs and user satisfaction
– Creating customer confidence in the product
– Understanding how and why the product creates value for the users
 Value turnover includes four building blocks: customer segments,
channels, customer relationships, and revenue stream. These building
blocks describe how the organization disseminates its value
proposition to the customers and how the customers generate
revenues for the company.
Value generation includes the building blocks of key partners, key
activities, key resources, and cost structures. More specifically, it
describes what is needed in terms of resources, activities, and partners
to create the value proposition and the costs associated with these
activities. Value generation should also specify—either directly or
indirectly—the value model (see ► Chap. 8) used by the company and
whether the platform is single-sided or multisided (► Chap. 10).
The BMC of specific organizations is a thorough description of each
of the nine building blocks and the relationships between them. ◘
Figure 19.2 shows an example of these relationships.

Fig. 19.2 Relationships between the building blocks in the BMC. (Authors’ own figure)

The steps in the analysis are as follows (the numbered list refers to
the numbers in ◘ Fig. 19.2):
1. The analysis starts with defining the set of value propositions the
organization delivers to the customers. This includes the goods and  
services that the organization delivers and the value and benefits
they may have for the customers.
2.
Identify the various customer segments and the value proposition  
each segment receives.
3.
Describe how the product is provided to the identified customer  
segments through specific channels, for example, over the Internet
or by postal services.
4.
Describe how customers directly and indirectly generate the  
revenue stream for the organization, also including customers or
users receiving the product for free (i.e., those with zero average
return per user (zero ARPU)) because their contribution may be
significant through indirect revenue channels.
5.
Identify the relationships between the organization and the  
customers.
6.
Identify the set of key activities that are required to create and  
support one or several value propositions.
7.
Identify the key resources required for these activities and for  
other purposes.
8.
Determine the costs associated with key activities and key  
resources and the overall cost structure of the organization.
9.
Identify strategic relationships with one or several key partners to  
support or contribute to the value proposition, key activities, and
key resources.
The BMC is used to model high-level abstractions of an
organization’s business operations. For two different organizations
doing business in the same business domain, there might be small
differences in the resulting BMC. However, even a small difference in
the BMC of two organizations—for example, if one of them is using the
core competencies in a smarter way—may render the business
operations of the two organizations radically different. If the BMCs of
the two organizations are identical, there is a motive for the
organizations to reconsider the BMC to create significant differences in
their business operations to distinguish it from that of the competitors.
To gain a strategic advantage, it is often enough to focus on one
specific building block in the BMC. It is seldom necessary to redesign
the business model completely to differentiate itself from the
competitors. This may be a critical issue if the value proposition offered
by a company is a commodity. In that case, price may be the only
parameter that distinguishes the competitors.
Successful business models may change over time—what turned
out to be a successful business model 5 years ago may not be a
successful business model today. The reasons may be technological
development, altered competition arena, new user demands, and
different market behavior. Technological developments may render a
business model obsolete.

19.3 The Stakeholder Relationship Model

The relationships between an organization and its stakeholders are
important aspects of its business operations. The stakeholder
relationship model (SRM) identifies key stakeholders engaged in the
organization’s business model and the interactions that the
organization has with these stakeholders. Stakeholders in the SRM are
the organization itself, customer segments, and key partners. The
customers and the key partners are identified and described in the BMC
as explained above. Relationships defined in the SRM can, for instance,
be exchange of assets (e.g., services, value, or money), formal
agreements, network effects, or other dependencies between the
stakeholders. ◘ Figure 19.3 shows the notations used to visually model
the SRM.
Fig. 19.3 Notations used in the stakeholder relationship model. (Authors’ own figure)

One key relationship between stakeholders is network effects (see

► Chap. 9). Network effects can either be positive or negative. The SRM
models three different kinds of network effects between two
stakeholders, A and B: (+/+), (−/−), and (+/−). The (+/+) network effect
means that stakeholder A induces a positive network effect on
stakeholder B and vice versa. Users in a telephone network have, for
example, a positive network effect on the other users in the network.
The (−/−) network effect implies that stakeholder A has a negative
network effect on stakeholder B and vice versa. One example is highway
traffic in which each car has a negative network effect on other cars on
the same road because of potential traffic congestion and increased
probability of accidents. The (+/−) network effect means that
stakeholder A has a positive effect on stakeholder B, but stakeholder B
has a negative effect on stakeholder A. One example is commercials on
television: the number of viewers has a positive network effect on
advertisers that want a large audience for the commercials, while
advertisements interrupting the flow of the program have a negative
network effect on the viewers.
The SRM supplements the BMC by visualizing the relationships
between the organization and other stakeholders. One important
purpose of the SRM, especially for businesses in the digital economy, is
that it illustrates the network effects that may modify the competitive
strength of the organization. Sometimes, these network effects depend
on each other and induce positive feedback that adds to the complexity
of the business model.

19.4 Digital Business Models

In a digital business model, the value proposition is related to a digital
good or service. Typical examples include the business models of
Facebook, Spotify, Twitter, Wikipedia, Microsoft Windows, and Skype.
The definition also includes the business models of sharing services—
such as Uber and Airbnb—since the platform that enables mediation
between buyers and sellers is digital. Platforms used for the sale of
tangible goods—such as Amazon and eBay—are also examples of
digital business models.
Every digital good or service is associated with a unique business
model. For multisided platforms, for example, Facebook, it may
sometimes be convenient to specify more than one business model.
Facebook may have one business model for attaching users and another
business model for capturing advertisers and other stakeholders
buying access to data Facebook collects about the users. In this case,
there is a strong relationship between the models because of strong
cross-side network effects.

Definition 19.3 Digital Business Model

In a digital business model, the value proposition is a digital good or
service.

► Case studies 19.1, 19.2, 19.3, 19.4, and 19.5 present five examples of
digital business models that have had a substantial impact on the
evolution of the digital economy: World of Warcraft
, Spotify, Facebook,
Wikipedia, and Airbnb
. These examples represent different types of
business models since they exploit the various fundamental properties
of the digital economy differently. ◘ Table 19.2 summarizes the
business models according to type and the most fundamental
properties exploited by the particular digital service. Note that the
business models presented in the case studies do not represent an
exhaustive list of all business models available for digital services.
Table 19.2 List of presented business models. (Authors’ compilation)

Digital Type of business model Payment method Fundamental

service properties exploited
World of Subscription All users pay a fee for using Zero marginal cost,
Warcraft the service network effects
Spotify Freemium Some users get the service Long tail, network
for free. Other users pay for effects
using the service
Facebook Ad-based free The social network service is MSP,
zero ARPU, zero
free. Advertisers pay for marginal cost,
user-targeted ads network effects
Wikipedia Commons-based peer No fees for readers and Public good, CBPP,
production authors. Revenues based on MSP, zero ARPU, zero
donations marginal cost
Airbnb Multisided digital Users pay for the mediation MSP,
network effects
mediation platform services provided by the
supporting non-digital platform
service

Case Study 19.1 World of Warcraft

Blizzard’s online game, World of Warcraft
(WoW), is an example of a
digital service based on the subscription
business model
. Here,
customers (or subscribers) pay a periodic (e.g., monthly or yearly)
fee for accessing the service. Without subscription, customers are
not allowed to access the service. If a customer terminates the
subscription, the customer also loses access to the service.
WoW is a massively multiplayer online game (MMOG) released in
2004. The game had over 12 million concurrent gamers
(subscribers) at its peak in 2010. WoW features a persistent 3D
world, where gamers can interact, solve quests, and perform tasks in
collaboration with other gamers. New content and upgrades are
continuously added to the game by Blizzard’s team of game
developers. Blizzard offers access to several servers—each featuring
a persistent world—to ensure load balancing and optimal
performance of the game. The persistent worlds are divided into
several geographical regions.
The BMC and the SRM of WoW are outlined in ◘ Figs. 19.4 and
19.5, respectively. The value proposition of WoW consists of giving
gamers access to the game and the persistent world’s content (1).
The only defined customer segment is gamers (2). The gamers pay
monthly subscription fees to access the game in addition to
purchasing the game itself. Both the subscription and the game
purchase are done via Blizzard’s battle.net website. The game can
also be purchased at a physical retail store. The main source of
income from the game is monthly subscriptions, which are
proportional to the number of active gamers (3). Another source of
income is the product sales from the WoW game itself.

Fig. 19.4 World of Warcraft

modeled using the BMC. (Authors’ own figure)
Fig. 19.5 World of Warcraft
modeled using the SRM. (Authors’ own figure)

The key activities needed to offer WoW is creation of content and

game development. These activities are the core activities for
Blizzard’s own development team and are essential for the value
proposition. In other words, the key resources are the game
developers themselves. In addition, Blizzard needs an infrastructure
of game servers to run the game. Both the game developers and the
server infrastructure are significant elements contributing to the
total cost of the company. Expenses related to developing the game
are, to a large degree, fixed and independent of the number of
gamers (4). Expenses related to the game infrastructure are partly
dependent on the number of gamers (5). The key partners are data
centers running persistent WoW worlds, ISPs providing high-speed
worldwide Internet access, and retail stores promoting and selling
the game.
Gamers exhibit positive direct same-side network effects on
other gamers—or, at least, on other gamers on the same server—as
shown in ◘ Fig. 19.5. Gamers can come together in teams—or
“clans”—to perform quests and tasks together, thereby increasing
the gaming experience. Gamers also provide in-game items and
content that other gamers may use. WoW Auction House is an
example of a place where gamers can sell and buy items found
during the game.
Key strategies for WoW are to acquire and retain customers who
will pay monthly subscriptions. This can be done by aggressive
marketing, offering a superior product, or exploiting the benefits of
network effects. WoW has exploited all these strategies. WoW is the
most successful MMOG developed so far, at least from a financial
point of view. Due to its success and impact in the gaming industry,
WoW has been a trendsetter of business models for online video
games.

Case Study 19.2 Spotify

Spotify is an example of a digital service that uses the freemium
business model
. Here, the digital service is offered to two different
consumer segments: One segment gets the service for free, while the
other segment pays for the service. The consumer segment that gets
the service for free is offered a simple—or “stripped down”—version
of the service. The consumer segment paying for the service is
offered access to all features of the service. Both consumer segments
are important in the business model since there are positive network
effects between them (Anderson, 2009). Spotify offers a music
streaming service either for free or for a monthly subscription fee.
The users receiving the service for free must listen to or view
advertisements, while those paying a subscription fee can listen to
music without interruptions.
Spotify’s business operation is modeled using the BMC and the
SRM as shown in ◘ Figs. 19.6 and 19.7, respectively. Spotify has two
value propositions (1): one for music streaming services and one for
advertisements. The subset of listeners paying for the service
contributes to the revenue of Spotify (2). This corresponds to about
90% of the revenue (2017) (Spotify’s revenues from 2012 to 2019,
by segment. Statista.). The other source of revenue—advertisements
—constituted about 10% of the income (2017). Key activities
include software development and content management. Software
development includes expanding, maintaining, and upgrading
software and infrastructure of servers and databases. Content
management is performed in close collaboration with the copyright
owners (the music industry). A key activity is to acquire the rights to
offer licensed music to the consumers (3). For these rights, Spotify
must pay royalties to copyright owners.

Fig. 19.6 Spotify modeled using the BMC. (Authors’ own figure)
Fig. 19.7 Spotify modeled using the SRM. (Authors’ own figure)

The two segments of paying and non-paying listeners induce

positive network effects on each other as shown in ◘ Fig. 19.7:
getting more free music listeners also increases the value for those
who pay for the service. This is so because a larger user base means
that Spotify can negotiate better deals with the music industry. More
importantly, with a large user base, Spotify will be in a stronger
position when competing with other music streaming services. Users
can also exchange playlists with one another.
The number of non-paying users induces positive indirect
network effects on the advertisers since more users means a
potentially larger audience for the advertisements. On the other
hand, advertisements—or rather the amount of advertisements—
induce negative indirect network effects on the users. This is
because advertisements are, for most people, an annoyance and
disruptions from using the digital service.
Spotify’s business model has had an effect on how people pay for
content on the Internet. It has also changed the size and operations
of the online content piracy industry.
Key strategies of Spotify—operating under the freemium
business model—are to minimize costs related to marketing,
customer acquisition, and customer care because most of Spotify’s
customers are non-paying customers generating only a small income
per customer. The foremost challenge for Spotify is to acquire and
keep customers and to communicate with them for the lowest cost
possible.

Case Study 19.3 Facebook

Facebook offers a social networking service in which users can
interact, socialize, share pictures and videos, play games, and use
other professional content. Facebook uses the ad-based free
business
model
. Here, users have free access to the digital service but must
give away control of personal data to Facebook and accept that they
will be exposed to advertisements. Revenue is generated by selling
targeted advertisement space based on data collected about the
users. Unlike the freemium business model, all users have access to a
complete and full version of the digital service. Facebook is the
world’s biggest social network service with more than two billion
monthly active users. Its international impact has influenced
business models worldwide in several other business sectors, as well
as motivated how people organize their social lives and spend time
on the Internet. Facebook is modeled using the BMC and the SRM in
◘ Figs. 19.8 and 19.9, respectively.
Fig. 19.8 Facebook modeled using the BMC. (Authors’ own figure)

Fig. 19.9 Facebook modeled using the SRM. (Authors’ own figure)

Facebook offers three value propositions to three different

customer segments (1): users involved in social interactions,
advertisers, and third-party content developers. Revenue is
generated from advertisements and service fees from third-party
content developers (2). About 90% of Facebook’s revenue is from
advertisements. Due to the large number of users, it is not possible
to establish a personalized customer relationship with every user.
Instead, Facebook relies on automated messages to nurture
relationships with the user segment. All customer segments access
the value proposition through the Facebook web page (► facebook.​
com) or the Facebook app (3).
Key activities are the software development of ► facebook.​com
and the Facebook app (4). Key resources are the infrastructure
hosting the Facebook service and the facilities storing and analyzing
personal and behavioral data collected about the users (5). The
stored user data is crucial for the business operations since it is used
to provide user-targeted advertisement space to marketers. This
data may also be sold to third parties for production of statistics and
other materials based on user behavior, for example, political
preferences, personality type, personal economy, and attitudes.
Trade and storage of such data was the key issue in the Cambridge
Analytica event in 2018 (Facebook value drops by $37bn amid
privacy backlash, 2018).
The major costs include salaries for software developers, service
management, and data storage and analysis (6). Key partners are
third-party content developers (7) who offer content (e.g., games)
directly to Facebook users.
As seen in ◘ Fig. 19.9, there are strong positive direct same-side
network effects in the user segment since gaining new users implies
that there are more opportunities for communication and
interaction in the social network. Content developers exhibit positive
direct network effects on users since gaining more users creates a
bigger market for content, while attaching more content providers
means that there is more available content for the users.
Advertisements, on the other hand, may have a negative network
effect on the users. For most users, advertisements are disturbing
and annoying and reduce the pleasure of social networking.
Finally, observe that Facebook offers the social interaction
service for free. Hence, the business model of Facebook is based on a
model where the average return per user (ARPU) is zero.
Case Study 19.4 Wikipedia
Wikipedia offers an online encyclopedia in several languages via its
website ► wikipedia.​org and the Wikipedia app for mobile devices.
The content of Wikipedia is created by thousands of contributors.
These contributors write new content on Wikipedia and edit,
correct, update, and quality-check content from other authors.
Wikipedia is used by millions of readers from all over the world.
Wikipedia is an example of the commons-based peer production
(CBPP) business model
. In CCBP, as explained in ► Chap. 7, the value
proposition is not produced by a firm but by a potentially large
number of people who, in a collaborative way, contribute to the
development of the service, normally, not receiving any financial
reward for their contribution. The reward may be in the form of
recognition, respect, or the satisfaction of having contributed to the
evolution of the content. CBPP requires effective mechanisms to
ensure collaboration between peers across distance, cultures, and
timespans since the collaborators may come from all over the world.
Recovering the costs of running and developing the service depends
on donations and public or private funding. The benefactors
providing these donations do not gain any direct revenue from the
digital service receiving the donations.
Wikipedia is modeled using the BMC in ◘ Fig. 19.10. The value
proposition is access to the online encyclopedia (1). Wikipedia is
available for free to users (readers) on ► wikipedia.​org and the
Wikipedia app (2). Hence, it is considered to be a public good.
Wikipedia is created by many writers who edit, create, and manage
Wikipedia articles through an online editing service (3). Key
partners are the benefactors who provide all the revenue needed to
run Wikipedia (4). Even though the content of Wikipedia is created
without any financial costs, there are costs of running the servers
and for covering the salaries of Wikipedia’s employees (5).
Wikipedia is owned by the nonprofit organization, Wikimedia
Foundation.
Fig. 19.10 Wikipedia modeled using the BMC. (Authors’ own figure)

The SRM of Wikipedia is shown in ◘ Fig. 19.11. There are three

stakeholders: benefactors, writers, and readers. All three
stakeholders have a relationship to Wikipedia—the writers develop
Wikipedia, the readers use Wikipedia, and the benefactors fund
Wikipedia. There are no relationships between these stakeholders
because:
– The authors are anonymous.
– Usually, several authors contribute to each article.
– The readers may also take the role of watchdogs monitoring the
quality of the content and, if necessary, correct it.
– The benefactors contribute because of the quality and correctness
of the encyclopedia and the importance the encyclopedia has on
society at large.
Fig. 19.11 Wikipedia modeled using the SRM. (Authors’ own figure)

Wikipedia is an example of a large digital service without any

form of relationships between the stakeholders. Hence, there are no
significant network effects associated with the various actors in the
business model.
The two main strategies of Wikipedia are to (1) develop and
maintain a platform that encourages writers to create and edit
content and (2) to ensure that the content is relevant, correct, and
important.

Case Study 19.5 Airbnb

Airbnb is an example of an organization using the multisided
platform
business model
(see ► Chap. 10). Airbnb offers a website
and a mobile app in which people may lease or rent houses,
apartments, or single rooms for shorter or longer stays. They have
also recently entered other businesses such as restaurant booking,
concert booking, and videos promoting different places around the
world. Hosts announce the availability of properties for rent on the
Airbnb website or app, often supplemented with photos, videos, and
describing text. Payment is done via the Airbnb website or app using
the Airbnb payment service. The host and guest arrange practical
details regarding the rental without involving Airbnb directly.
In this business model, the organization (Airbnb) offers a
platform for mediating between and connecting users belonging to
different groups (hosts and guests). Unlike the ad-based business
model, the multisided platform business model does not use ads to
generate revenue but instead receives revenue from transaction fees
levied when users exercise the services offered by the platform. In
the platform business model, the matching of users is often
motivated by the exchange of personal services (e.g., Airbnb and
Uber) or tangible goods (e.g., eBay). There are several variations of
the multisided platform business model; however, the key idea is
that a platform is employed for connecting different user groups. The
BMC the SRM of Airbnb is illustrated in ◘ Figs. 19.12 and 19.13,
respectively.

Fig. 19.12 Airbnb modeled using the BMC. (Authors’ own figure)
Fig. 19.13 Airbnb modeled using the SRM. (Authors’ own figure)

The main value proposition (1) of Airbnb is to offer a mediation

service between the two customer segments hosts and guests (2).
This is done on the website ► airbnb.​com or the Airbnb mobile app
(3), in which the guest must provide a valid name, email address,
telephone number, photo, and payment information. Airbnb gets its
revenue from a booking fee paid by both the guest and the host (4).
Key resources are the website ► airbnb.​com and the mobile app (5).
These sites must be user-friendly for both hosts and guests, have
high availability, and offer a high-quality service. Key partners are
people owning properties for lease (the hosts) (6). Hence, hosts are
both a key partner and a customer segment for Airbnb. This is
because hosts provide the content of the Airbnb site (properties for
rent), while at the same time, they contribute to revenues when a
guest books the host’s property for rent. A key activity is network-
building to build up a large base of available hosts.
The business model contains positive indirect network effects
between guests as seen in ◘ Fig. 19.13. Guests leave comments that
other guests may review before booking a property. There are also
positive direct network effects between guests and hosts—having
 more hosts means there are more choices available for the guests
and more guests implies a higher likelihood of bookings. There are
negative network effects between hosts since hosts compete for the
same guests. However, note that the network effects have local
significance only: available properties in Bangkok do not induce
positive network effects on users who are looking for a house in
Paris.
A key strategy for Airbnb is to ensure a healthy supply of
property in the areas Airbnb operates and to promote these
properties to attract enough guests. Another important aspect is to
ensure that guests trust the bookings made through Airbnb. For this
purpose, Airbnb offers a review system for both hosts and guests
that may be used to rise the quality of the services offered by Airbnb.

19.5 Conclusions
Business models are developed to identify key aspects of an
organization’s business operations. The models are used both to
describe the current state of the organization and to identify key
strategic actions required for taking the organization into the future. In
other words, the purpose of the model is to identify the current market
position of the organization, estimate how the market and the
technology will evolve, and on this basis, develop strategies for future
market positioning and revenue generation.
The business model canvas developed by Osterwalder is a simple
and effective tool to develop and analyze business models for
enterprises in the digital economy. These enterprises are often
multisided platforms where each business sector supported by the
platform is based on entirely different value proposition, production,
cost, and revenue models. The stakeholder relationship model
supplements the business model by identifying the type of relationship
that exists between the stakeholders and the impact these relationships
have on the business model. The principles have been illustrated for
five enterprises with entirely different underlying value proposition
models.
Questions
1.
Massive Open Online Course (MOOC
) was envisioned to disrupt  
the educational sector by offering virtually free and ubiquitous
teaching online.
(a)
Would you categorize MOOC as either a sustainable  
innovation or a disruptive innovation?
(b)
Design the business model of a company offering an MOOC  
(e.g., Coursera) using the business model canvas.
2.
Popcorn Time (now terminated by regulations)
and Netflix were  
two digital services with similar value propositions.
(a)
Use the BMC and the SRM to design the business model for  
Popcorn Time and Netflix.
(b)
What are the major differences in the resulting business  
models?
(c)
What is the major technological difference between  
Popcorn Time and Netflix?
3. Search the web and find answers to the following questions
concerning Spotify:  
(a)
How many of Spotify’s users subscribe to its premium  
service?
(b)
What is the largest cost of running Spotify?  
(c)
Does Spotify generate profits?  
(d)
Which companies are Spotify’s main competitors?  
(e)
What is Spotify’s market share in the music streaming  
 industry?

(f) Based on your answers, identify the major strategic

4.
challenge for Spotify?  
Wikipedia  
(a)
Why does Wikipedia rely on donations and not  
advertisements to provide revenue for its business
operations?
(b)
Is it possible for Wikipedia to use the subscription-based  
business model?

Answers
1.
MOOC (see Wikipedia article)  
(a)
Currently, MOOC looks more like a sustainable innovation  
than a disruptive innovation. This is because it has not
changed the performance metrics or value chain in the
education business. It is a sustainable innovation because
it has added to the current selection of teaching methods
that teachers may employ. However, MOOC has the
potential to become a disruptive innovation in the future.
However, predicting such an event is speculative at best.
(b)
A company offering an MOOC is a value network and an  
MSP. It connects two user groups: teachers and students. A
key insight in the business model of an MOOC is that the
teacher user group is producing the key resource of the
MOOC—the teaching material. Another insight is related to
the revenue stream. Should MOOC be free of charge or
should there be a fee for the students? If there is a fee, how
should this fee be distributed between the provider of the
MOOC and the teachers? (◘ Fig. 19.14)
P Ti d N fli
2. Popcorn Time and Netflix
(a)
Popcorn Time and Netflix are both providers of online   
streaming media (e.g., video, series, and movies). The
BMCs are as shown in ◘ Figs. 19.15 and 19.16.
(b)
The major business difference is that Popcorn Time is free  
and has a disputed legality, as users uploading and
distributing content in many cases (but not all) do not have
the copyright owners’ permission. Netflix, however, is a
paid service and legal, as Netflix has bought copyright
content in addition to producing its own content.
(c)
The major technological difference between Popcorn Time  
and Netflix is that the former is based on a peer-to-peer
architecture, while the latter is based on a server-client
architecture. On Popcorn Time, content is streamed from a
network of users (BitTorrent technology). On Netflix,
content is streamed from a server owned by Netflix. This
results in strong network effects between users of Popcorn
Time and non-existing network effects between users of
Netflix.
3. Data for Spotify December 2020 retrieved from Spotify’s
homepage, Statista, and Wikipedia:  
(a)
155 million premium users.  
(b)
Major costs are licenses to the music industry.  
(c)
No—losses were 186 million euros in 2019.  
(d)
The three biggest competitors are Apple Music, Amazon  
Music, and TenCent.
(e)
The market share was 32%.  
(f)
The major challenge is how to become profitable.  
4.
Wikipedia article about Wikipedia:  
(a)
Wikipedia believes that ads are annoying and distracting  
for the users, will influence the neutrality and threaten the
credibility of the content, and may generate conflicts of
interest between stakeholders that lead to censorship of
the content.
(b)
Currently, Wikipedia cannot use the subscription-based  
business model. This is because Wikipedia uses the GNU
Free Documentation License.

Fig. 19.14 MOOC modeled using the BMC

Fig. 19.15 Popcorn Time modeled using the BMC

Fig. 19.16 Netflix modeled using the BMC

References
Anderson, C. (2009). Free: The future of a radical Price. Hyperion.

Christensen, C. (1997). The innovator’s dilemma: When new technologies cause great firms to
fail. Harvard Business Scholl Press.

Facebook value drops by $37bn amid privacy backlash. BBC News. March 19, 2018.

Osterwalder, A., & Pigneur, Y. (2010). Business model generation: A handbook for visionaries,
game changers, and challengers. Wiley.

Ovens. (2015, January). What is a business model? Harvard Business Review.

Further Reading
Anderson, C. (2009). Free: The past and future of a radical Price. Hyperion.

Granstrand, O. (2016). Industrial innovation economics and intellectual property. Svenska

Kulturkompaniet.

Osterwalder, A. (2010). Business model generation. Wiley.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_20

20. Big Data Economics

Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Artificial intelligence – Machine learning – Expert systems –

Data mining

Learning Objectives
After completing this chapter, you should be able to:
– Understand the basic concepts related to big data.
– Identify business opportunities offered by the big data
technology.
– Explain how big data can be abused.

20.1 Zettabyte Era

Generation, processing, distribution, and storage of data have
undergone major transformations during the last 30 years (Hilbert &
López, 2011; Cisco Annual Internet Report (2018–2023) White Paper.
Published March 9, 2020) (the data below are based on raw data from
these two sources).

20.1.1 Generation
In 1993, only 3% of all data stored was digital; the rest was stored on
analog media such as books, cassettes, and video tapes. In 2007, this
number had increased to 84%. Today, almost all data is available in a
digital format. The main reasons are that most of the analog data from
earlier times has been converted to digital data and, more important,
that almost all data generated after 1990 has been produced directly in
a digital format. The evolution is shown in ◘ Fig. 20.1. The total amount
of annually generated data is shown in ◘ Fig. 20.2. The amount of data
generated in 2019 is estimated to be 43 zettabytes, and it is predicted
that this number will increase to 175 zettabytes in 2025 if the amount
of data produced continues to increase by 23% per year. One zettabyte
is one trillion gigabytes or 1021 bytes. Data is generated by four major
sources: imaging (e.g., medical imaging and surveillance cameras),
entertainment (e.g., television and radio shows, videos, podcasts, social
media, and video games), manufacturing and administration (e.g.,
automation, Internet of Things, and word processing), and voice
(mobile phones and VoIP). Not all this information is stored; for
example, most telephone conversations are not stored (Reinsel et al.,
2018).
Fig. 20.1 Percentage digitally stored data. (Authors’ own figure)
Fig. 20.2 Annually generated data. (Authors’ own figure)

20.1.2 Processing
The doubling time for the processing capacity in terms of instructions
processed per second (IPS) is estimated to be 1 year and 2 months for
general purpose computers and 10 months for application-specific
computers (e.g., supercomputers). The processing capacity cannot be
determined exactly since it is a complex function of variables such as
memory capacity and organization, clock rate, programming language,
computer architecture, and operating system. The above numbers are
therefore crude estimates of the doubling time for the processing
capacity.

20.1.3 Distribution
In 1993, more than 30% of the telecommunications networks and
almost all broadcast networks were still analog even though the
digitization of these networks started 20 years earlier. In 2020, the
digitization is complete, and all information sent over the
telecommunications networks is digital. Since 2000, the Internet has
become the dominant carrier of information. The doubling time for the
capacity of the Internet has been 1 year and 8 months during the last
20 years as shown in ◘ Fig. 20.3. In 2016, one zettabyte of data was
transmitted over the Internet. The largest contributor to the traffic
growth is video streaming.

Fig. 20.3 Capacity of the Internet. (Authors’ own figure)

20.1.4 Storage
The amount of stored data has increased by about 30% per year. This
corresponds to a doubling time of about 2 years and 4 months. The
amount of data stored surpassed one zettabyte in 2012 and is
estimated to reach almost 20 zettabytes in 2020.
During the last 10 years, the amount of data generated, stored, and
distributed has surpassed one zettabyte (1021 bytes), and computer
science has entered a new historic period, the Zettabyte Era.
The Zettabyte Era has laid the basis for the big data technology—the
systematic analysis of huge data sets, usually too large and
unstructured to be handled by traditional methods. This may be viewed
as a possible, though vague, definition of big data technology. Big data
economics may, in a similar vein, be defined as follows.

Definition 20.1 Big Data Economics

Big data economics is the study of how big data can be turned into
economic value, where the key value proposition is systematic
processing of digital data to detect hidden information that can be
used for business purposes.

Big data usually refers to advanced data analysis methods to determine

user behavior, to uncover patterns in large data sets, and to extract
particular information from unstructured data. The term does not refer
to the size of the data set itself since even large data sets may contain
information that can be extracted by simple methods (e.g., identities
and addresses of persons stored in national registers). On the other
hand, the data set must be large in order to contain information that is
hidden by structural complexity and, therefore, requires big data
analysis techniques to be uncovered. ► Box 20.1 lists some of these
technologies.

Box 20.1 Artificial Intelligence, Machine Learning, Expert

Systems, and Data Mining
Big data analytics often requires complex computational methods
and draws on methods developed in other fields of computer science.
 Artificial intelligence (AI) is “defined as a system’s ability to
correctly interpret external data, to learn from such data, and to use
those learnings to achieve specific goals and tasks through flexible
adaptation” (Kaplan & Haenline, 2019). Big data analysis is not a
subfield of AI but may apply methods and tools developed for AI, for
example, advanced search algorithms to identify hidden information,
image analysis for face recognition, learning algorithms predicting
customer behavior that can be used for targeting advertisements and
individual pricing, trading algorithms for the stock market, trend
predictions, and several other technologies.
Machine learning uses computer algorithms that are
automatically updated and modified as new information and
experience is gathered about the family of problems to be solved.
Machine learning is an efficient method in cases where the
algorithms are too complex or infeasible for standard algorithm
design, for example, for spam filters for emails and navigators for
trucks in automated warehouses.
Expert systems consist of a knowledge base and a set of inference
algorithms. The knowledge base is continuously updated by external
input and by internal machine learning algorithms where results of
earlier predictions are fed back to the knowledge base. The inference
algorithms consist of if-then rules to estimate the consequences of a
decision (scenario analysis and decisions under uncertainty). Expert
system technology may be used to manage business operations and
customer relations.
Data mining refers to large-scale data analysis methods to
discover patterns and dependencies in complex data sets, in
particular, unknown patterns and dependencies. Among the methods
used are AI, machine learning, statistics, mathematical inference, and
database management. One particularly challenging problem in
statistical analysis of complex data sets is causality. The purpose of
causality analysis is threefold (Pearl, 2009):
– To determine if two variables actually depend on each other and, if
so, to determine which of them is the cause and which is the effect
– To determine if two independent variables are correlated because
there is a common cause that the two variables are correlated
 – To determine if the variables are independent and the correlation
between them is accidental
To make things even worse, correlation implies that the two
variables are linearly related. If this is not the case, the variables may
by strongly related, but the correlation between them is zero; for
example, if one of the variables increases as the square of the other
variable, then the variable are strongly related but the correlation
between them is zero!

20.2 Characteristics of Big Data

Big data are defined by the following characteristics (see ◘ Fig. 20.4)
referred to as the four Vs:
– Volume, referring to the amount of digital data available for analysis.
– Variety, referring to the richness of data categories. During the
Zettabyte Era, stored data has evolved from being mainly structured
data—data that is organized and adheres to a specific format—to
unstructured data such as fusion of text, location data, video, images,
and social media activity. Such complex data structures cannot be
processed or analyzed by simple analytical tools.
– Velocity, referring to the speed by which data is generated and
processed. Big data is usually produced in continuous processes.
Some of this data is captured, processed, and published in real time.
– Veracity, referring to the exactness of the data. The accuracy of the
data analysis depends on biases, noise, inaccuracies, and
irregularities in the data set. Insofar as possible, the data analysis
should take such anomalies into account and estimate the validity of
the result of the analysis.
Fig. 20.4 Growth in key characteristics of big data. (Authors’ own figure)

The huge volumes of data are generated by both people and

machines. Data produced by people include photos, text, videos,
movies, music, professional content, video conferencing, and chat logs.
Data produced by machines (M2M) include sensor data (e.g.,
measurement of environmental data, health monitoring, and intelligent
traffic management), medical images (in particular MRI), videos from
surveillance cameras, and system updates and configurations. The
annual compound growth rates for the various segments of M2M
communication are expected to be between 10% and 50% from 2018
to 2023 (Holst, 2018).
Big data analytics includes tools and techniques to convert vast
amounts of raw data into meaningful information that can be used as
tradable goods or for optimization or personalization of services
offered to users and customers. The input to the big data analytics
algorithms is the vast amount of digital data that is collected about
users, processes, and events. The output is information used in
marketing, business planning, behavioral control, trend analysis, and
statistics. Unstructured data turned into meaningful information is the
basis for the value propositions of the big data services.
An example of a computer facility designed for big data storage and
processing is shown in ► Box 20.2.

Box 20.2 The Lefdal Mine Datacenter

Handling the enormous amount of data generated in the Zettabyte
Era requires huge installations of processors, storage devices, and
communication equipment, sometimes covering several football
fields. These datacenters consume massive amounts of electricity,
especially for cooling of equipment to avoid overheating. Public
awareness, environmental impact, cost efficiency, and social policies
require that such datacenters operate with high energy efficiency
and at the lowest cost possible.
The Lefdal Mine Datacenter is a company offering storage
capacity and space for installation of data equipment in an
abandoned mine deep into the mountains in the western part of
Norway. The mine offers a cool environment reducing the
requirement for artificial cooling. Additional cooling is supplied by
cold seawater from the fjord nearby. Moreover, the center is powered
only by renewable hydroelectric power from the Norwegian power
grid. Lefdal Mine Datacenter claims to be the “greenest” datacenter in
Europe (Homepage Lefdal Mine Datacenter).
High-speed Internet communication with mainland Europe and
the UK is by fiber cables from the Norwegian coast to the UK and the
Netherlands.
The Lefdal Mine Datacenter is not the only cave or mine that has
been used as datacenter. Similar centers exist in, for example, Finland
 (Sverdlik, 2015) and the USA (Abandoned mines have a future as
data centres, 2014).

20.3 Use of Big Data

Data abundance refers to the huge amount of data available today. Data
has become abundant and widespread to such a degree that it opens up
for completely new ways of doing business, including development of
new products, building up markets for these products, and finding new
ways to earn money. Several enterprises and organizations must realize
that their business models must be organized in radically new ways to
keep up with the evolution. Data management and analytics have also
become completely new industries, offering huge computer resources
for data management and specialized expertise in big data
technologies.
Some of the new opportunities offered by big data are described
next. In ► Sect. 20.4, we will come back to questions concerning ethics,
governmental control, and violation of personal integrity and privacy.

20.3.1 Marketing
Marketing and advertising are among the biggest applications of big
data. By sifting through the enormous amounts of information users
place on social media, inspecting how mobile apps are used, monitoring
web search, and recording bank card transactions, patterns may be
found that can be used in targeted and personalized marketing or to
promote sales in new ways.

20.3.2 Health Care

The most important application of big data in health care is computer-
aided diagnostics. This includes analysis of enormous data sets to
match symptoms with possible diseases and for interpretation of X-ray,
MR, and ultrasound images. Big data analytics is also important in
medical research to handle complex and enormous data sets generated
in research projects and for meta-analysis where data and results from
several independent studies are combined.

20.3.3 Algorithmic Financial Trading

Value and availability of stocks and securities in the stock market and
currency exchange rates in the foreign exchange market (FOREX)
fluctuate so rapidly that human traders are not able to follow them in
real time. On the other hand, traders may earn much money from these
markets if they are able to react to the fluctuations rather than to time
averages. Algorithmic trading is based on fast computer algorithms that
can react to changes in the market in microseconds and sell or buy
stocks or currencies at optimal instances. The technology requires
algorithms that can sift through large amounts of data and make
trading decisions within microseconds based on patterns in these data.

20.3.4 Government and Public Services

Governments may use big data to prepare more extensive national
statistics by combining information from public and private databases.
Big data may also be used to monitor and improve public services.
Fingerprints, retina scans, face recognition data, DNA profiles, and
other biometric data of own inhabitants and foreigners visiting the
country are also stored. These and other data may be used by
governments for social control of the population and monitor or expel
unwanted visitors.

20.3.5 Insurance
The insurance industry uses big data to predict variations in life
expectancy, health costs, and cost of natural disasters and personal
accidents using information from public and private databases.

20.3.6 Retailers
Big department stores collect enormous amounts of data about their
customers. If goods are paid by cash, the customer is anonymous but
not if bank and membership cards are used. In the latter case, the
retailer may use the information for personalized marketing. The
retailers also use the information for logistics and other administrative
purposes.

20.3.7 Data Brokers

Data brokers (or information broker) receive or buy information from
various suppliers, for example, social networks, retailers, influencers,
app owners, and so on. This may be unprocessed raw data or data that
has been processed to some degree. The data broker may then process
the data further and sell processed data to other organizations on
demand or in the open market.

20.3.8 Electronic Media

Social media, newspapers, journals, and the TV industry use data they
collect about the customers to produce targeted advertisements,
editorials, and articles. They may also use the data to detect changes in
user behavior, increase customer loyalty, and prevent churning. They
may also sell this information to marketers (see above).

20.3.9 Science
Big data analytics is a key feature of several of the world’s largest
scientific experiments, for example, the Large Hadron Collider at CERN
in Geneva, Switzerland, correlated gravity wave detectors in the USA,
Japan, and Italy, neutrino detectors, and big arrays of astronomical
radio telescopes. In all these cases, the problem is to sift through
enormous amounts of data to detect rare events that may predict new
physics and provide knew knowledge in the fields of astronomy and
cosmology. Big data is also used in sport sciences to determine the
effect of training, diet, and body functions measured by sensors in or
attached to the body.

20.3.10 Data Illiteracy

Big data analytics and artificial intelligence are highly specialized
disciplines taught in advanced courses at several universities. Both
fields require insight into highly specialized technologies and advanced
mathematical methods. The technology and its applications are not
taught at business schools. One concern is, therefore, that the benefits
of big data technologies often are neglected by the management
because of data illiteracy.
An example of how big data can be exploited to create several new
business opportunities for a media company is described in ► Box
20.3.
One problem associated with big data is lack of knowledge about
how the data can be used in decision-making. Many decision makers
and managers lack data literacy and do not understand how the
enormous amount of data the firm is collecting about itself and its
customers can be used to improve the business. Big data analytics is so
new and complicated that it is not taught in management courses and
at business schools. Investments in big data may then be useless if the
management does not understand how to use these data in decision-
making. On the other hand, the output from a big data analysis may be
flawed or irrelevant so that decision making cannot be based on the
data only; the decision maker must also use other knowledge and guts
feelings as a supplement to avoid fatal mistakes (Vigen, 2015). One
particularly important observation is that in huge data sets spanning
over several categories of observed data, it is likely to find
coincidentally correlated data. Obviously, this may lead to wrong and,
sometimes, disastrously conclusions. An eye opening and amusing
account of such coincidences is found in (Shah et al., 2012).
Decision making was easier before the Zettabyte Era because the
business processes then depended only on simple and easily
comprehensible data structures.

Box 20.3 Netflix and Big Data

Big data is an essential commodity for Netflix in its business
operations. Netflix gathers and stores data from its over 180 million
users and uses this data to discover user behavior and viewer
patterns. The data collected is detailed and creates a comprehensive
profile of the subscribers, for example, the time it takes the user to
finish viewing a movie or a show, devices used to view the show,
pauses taken during the show and whether the viewer continues to
watch the show after the pause, and which days of the week and time
of the day the user is active.
Netflix uses big data analytics to process this data and turn it into
useful information. The information is used to recommend movies
and TV shows to the users, matching their preferences as accurately
as possible. Over 75% of movies and TV shows the user watches are
based on recommendations by Netflix. Netflix also uses the
information to decide which new content to produce. Observing that
exceptionally many customers are viewing a particular series may
result in the production of new series in the same genre.
 Netflix also uses this information for marketing purposes, e.g., by
attaching different marketing trailers to the film or the show that are
tailored to match the interests of different user groups.
The recommendation algorithm used by Netflix does not require
any direct input from the subscribers. It is only based on information
collected from the Netflix servers. The economic value of
recommendation algorithms must not be underestimated. They may
reduce customer churn, increase the value of Netflix as a marketing
arena, and boost user satisfaction. Netflix may be regarded both as a
data analytics company and as a media company (Sivasubramian,
2020).

20.4 Abuse of Big Data

Big data often uses personal data. Because personal data may be
sensitive and contain private information that the subject do not want
to share, there are several legal frameworks that big data systems need
to adhere to. One such legal framework is the European General Data
Protection Regulation (GDPR), which came into force in the European
Union and the EEC in 2018. Other countries, including several US states
and many Asian countries, have lately adopted similar laws. Such
regulations limit the amount of personal data that may be harvested.
On the one hand, this in turn limits the value of big data, since some
useful data may not be collected due to legal issues.
United Nations Declaration of Human Rights Article 12 states: “No
one shall be subjected to arbitrary interference with his privacy, family,
home or correspondence, nor to attacks upon his honor and reputation.
Everyone has the right to the protection of the law against such
interference or attacks.” One of the key problems of big data is that it so
easy to violate this rule and so difficult to prosecute those who do so.
One particular problem is that the technology advanced very rapidly,
while the production of new laws protecting the users progresses very
slowly. Ownership of personal data is in itself a political question. In
China, the government claims that it owns all data about the
inhabitants.
Despite these regulations and legislations, personal data is used for
purposes that may be unethical or against the interests of the majority
of the population.
Some examples of misuse of big data are presented next.

20.4.1 Clandestine Operations

Big data is used by intelligence and security agencies for collection and
analysis of clandestine information. One example is the NSA-led PRISM
program, where the US government collects data from various Internet
service providers in the USA and by intercepting all Internet traffic into
the USA. A large part of the world’s Internet traffic passes through the
USA and is, therefore, the best source there is for collection of
clandestine information. Close cooperation with the Government
Communications Headquarters (GCHQ) in the UK also provides them
with clandestine information that GCHQ collect by intercepting Internet
traffic into the UK. The program is secret, but the existence of it was
leaked by Edward Snowden in 2013 and published in The Washington
Post and The Guardian.
Most of the large companies in the digital economy have provided
input to the program, including Microsoft, Google, Facebook, Apple, and
YouTube, just to mention some of the biggest contributors.
Another related interception program is he ECHELON project
operated by the USA, the UK, Canada, Australia, and New Zealand for
interception of satellite traffic, telephone traffic, and the Internet
(Gerhard, 2001). Its existence was disclosed by The Guardian in 2001:
“What is Echelon? A global network of electronic spy stations that can
eavesdrop on telephones, faxes and computers. It can even track bank
accounts. … Officially, however, Echelon doesn’t exist. Although
evidence of Echelon has been growing since the mid-1990s, America
flatly denies that it exists, while the UK government’s responses to
questions about the system are evasive” (Perrone, 2001).
The biggest problem that programs like PRISM and ECHELON
generate is that they are not under democratic control and can be
misused by the government to control and manipulate the population.
There are two major types of information that is intercepted and
stored: the content of messages and conversations and the metadata
associated with the messages. These concepts are explained in ► Box
20.4.
Box 20.4 Metadata, Content, and Privacy
Three concepts that are important in the context of data protection
are metadata, content, and privacy. In clandestine data collection, the
purpose is to find out as much as possible about the message that is
intercepted. This includes not only access to the plaintext data
contained in the body of the message but also harvesting as much
metadata about the data traffic as possible.
Metadata (or “data about the data”) includes identities of the
sender/origin and the receiver/destination, URLs identifying the
type of content, type of message (WWW message, email, file transfer,
VoIP, streaming service, etc.), protocol details (IP, UDP, TCP, and
tunneling headers, service initiation protocols, encryption method,
etc.), length of the message, and the time the message was sent or
intercepted. Even if the content of data cannot be read, the metadata
may provide the secret services with information from which they
may infer political, criminal, or other activities by using artificial
intelligence or big data algorithms for discovering patterns in the
data steam.
Content is the information contained in the data field of the
message, for example, the text in an email, bank account details and
the amount transferred in a bank transaction, search words in a web
search, and the content of a file.
Privacy is the act of hiding the information. This includes
encryption of information stored on hard discs and information sent
over the Internet. The most common encryption method used to hide
the content of Internet messages is Transport Layer Security (TLS)
encrypting the body of the message (e.g., the https protocol used for
protection of web messages). Better protection is provided by IPsec
used in Private Virtual Networks supporting secured
communications within and between geographically distributed
organizations. Metadata such as addresses and protocol types are not
protected by these methods.
Addresses may be kept secret by using the onion router (Tor)
where the address of the sender is anonymized. This reduces the
value of the metadata collected from such messages because it is not
possible to correlate sender and receiver of the messages. Tor is also
used in the dark web together with encryption to hide information
 and make transactions untraceable. The dark web is used by
terrorists, criminals, and hackers and for other illegal purposes, as
well as for legal purposes such as protection against industrial
espionage. If it is observed that frequent interactions take place
between two companies during a short period, a competitor may
conclude that a new business relationship (e.g., a merger) is planned.
Note that information may be sent in plaintext in local networks
and only be encrypted when sent on the open Internet, so that
tapping the information is still possible. This possibility is sometimes
used (or misused) by the management of some firms to control that
the staff is not using the Internet for private purposes.

20.4.2 Social Control

China’s social credit system is a governmental application of big data.
The Chinese government gathers information about its inhabitants
using a vast number of surveillance cameras, information retrieved
from online Internet activity, and financial transactions. The data is
used to calculate a credit score ranging from 0 to 1000 for every
Chinese individual. This score is used to determine whether an
individual will get access to particular services, worthy of getting loans,
or enjoy particular benefits. It is also used for punishment such as
exclusion from high prestige work, only getting slow internet access,
exclusion from certain schools, travel restrictions, or included in public
blacklists. The system is designed to promote “good” behavior in the
population.

20.4.3 Tracking People

Several companies have developed camera surveillance systems for
real-time face recognition (See for example the homepage of
Innovatrics). The technology is common in China where surveillance
cameras on streets, railway stations, airports, shops, and many other
sites are used together with face recognition algorithms to identify
people (A lawsuit against face scans in China could have big
consequences, 2019). This is part of the social credit system described
above.
Several mobile apps include GPS location data in the messages they
send to the server. To meet regulatory requirements, these data are not
coupled to the identity of the mobile user and are therefore claimed to
be anonymized. Tamoco is a UK-based company buying location
information from app owners and reselling it to customers (See the
official homepage of Tamoco). Coca Cola, Uber, and Nestlé are among
their customers. This allows the customer to follow the movements of a
particular mobile phone without knowing the identity of the user of the
phone. May 9, 2020, the Norwegian Broadcasting Corporation showed
how easy it is to use the anonymized data from Tamoco to identify a
person: determine where the mobile is usually located during the night
and during the day, and the owner is easily identified (► https://​www.​
nrk.​no/​norge/​xl/​avslort-av-mobilen-1.​14911685).

20.4.4 Violation of Personal Integrity

The users of social media and other services offered over the Internet
usually do not know what data the provider of the service collects
about them and for what purposes the provider is using this
information. Most of the revenues of the application service providers
are, in fact, generated by the information the provider has been able to
collect about their customers. One particular problem is that if the user
does not allow that data is collected and used for other commercial
purposes, they will be denied receiving the service. One case where
regulations help is the use of cookies. The service provider must show
the content even if you do not accept the use of cookies, though there
are still providers who deny access if you do not return the cookie, let
you only view part of the content, or do not remove the text of the
request so that it masks three quarter of the screen making it difficult
to read the content.
One of the most quoted examples of misuse of personal data is the
Cambridge Analytica scandal in 2018, where Cambridge Analytica
collected information from Facebook which they attempted to sell to
the Republican presidential campaign.

20.5 Conclusions
Nobel Laureate Ronald Coase once said: “if you torture the data long
enough, it will confess to anything” (Wiktionary). This is indeed a
truism for big data analytics. Treated without caution and skepticism,
big data analytics may lead to wrong—and sometimes disastrous—
decisions. Handled with care and expertise, big data is a formidable
competitive tool in the digital economy improving customer
satisfaction and perfecting sales and marketing operations by precisely
targeted information bulletins and ads.
The challenge is that the firm must possess deep knowledge in
advanced data management tools such as artificial intelligence,
machine learning, expert systems, and data mining. The result may be
that the company is not able to discover and utilize the huge amount of
data it may possess about its business operations and customers.
Big data offers big opportunities in several sectors, for example, in
health care, marketing, digital service provision, statistics, and
management of public services. Large amount of data about people is
collected by government bodies, both public and clandestine, by
intercepting Internet traffic; receiving data captured by social media
providers, telecommunications operators, and application providers;
storing information received from surveillance cameras; and storing
biometric information about inhabitants and visitors. This information
is used for crime prevention, criminal investigation, and antiterrorism.
The same data may be misused for social control of the population and
for identifying, tracking, and harassing dissidents and political
opponents.

Questions
1.
Is causation and correlation the same thing? Explain.  
2.
From which sources do data brokers collect information? Hint:  
see, for example, ProPublica (Everything We Know About What
Data Brokers Know About You) and Clearcode (What Is a Data
Broker and How Does It Work?).

Answers
1. No. That two events are causally related means that one event is
caused by the other event. That two events are correlated means  
that a linear statistical relationship exists between them (e.g.,
 both increase at statistically proportionate rates—or one
increases, while the other decreases at a proportionate rate).
The events may be correlated because the events are causally
related, both are caused by a third event and not causally related
themselves, or they are entirely unrelated but varies in the same
way (spurious correlation). If the relationship between two
variables is nonlinear, then the correlation between them is zero
so that causation does not imply correlation.
2.
The data brokers may buy or retrieve information from several  
resources, for example:
– Publicly available information (e.g., birth certificates, criminal
registers, taxation registers)
– Webpages mentioning the person
– Loyalty cards for shops, hotels, etc.
– Membership lists of organizations
– App owners
– Dealers of products that can be connected to data about the
customer (e.g., car dealers and realtors)
– Media providers
– Search engines
– Telecommunications operators
– Other data brokers

References
A lawsuit against face scans in China could have big consequences. The Economist, November 9,
2019.

Abandoned mines have a future as data centres. Aggregate Research, September 18, 2014.

For the story in Norwegian, follow the link https://​www.​nrk.​no/​norge/​xl/​avslort-av-mobilen-

1.​14911685

Gerhard, S. (2001). On the existence of a global system for the interception of private and
commercial communications (ECHELON interception system) (2001/2098(INI)).

Hilbert, M., & López, P. (2011). The World’s technological capacity to store, communicate, and
compute information. Science, 332(60), 60–65.
Holst, A. (2018, October 17). M2M (machine-to-machine) – Statistics & Facts. Statista.

Kaplan, A., & Haenline, M. (2019). Siri, Siri, in my hand: Who’s the fairest in the land? On the
interpretations, illustrations, and implications of artificial intelligence. Business Horizons, 62(1),
15–25.

Pearl, J. (2009). Causal inference in statistics: An overview. Statistic Survey, 3.

Perrone, J.. (2001, May 29). The Echelon spy system. The Guardian.

Reinsel, D., Gantz, J., & Rydning, J. (2018). The digitization of the world: From edge to core.
International Data Corporation.

Shah, S., Horne, A., & Campellá, J. (2012). Good data won’t guarantee good decisions. Harvard
Business Review.

Sivasubramian, B. (2020, August 7). How Netflix became $100 billion company using data
science. Analytics Vidhya.

Sverdlik, Y.. (2015, July 1). Start-up to build underground data Center in Finland. Data Center
World.

Vigen, T. (2015). Spurious correlations. Hachetts Books.

Further Reading
Ford, M. (2017). The rise of the robots. Oneworld Publications.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_21

21. Net Neutrality
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Device neutrality – Search neutrality – Zero-rating

Learning Objectives
After completing this chapter, you should be able to:
– Explain why net neutrality implies that ISPs must treat all
communication on the Internet equally and without any kind
of discrimination.
– Explain how net neutrality promotes innovation, prompts
competition, and supports free exchange of information on
the Internet.
– Identify reasons why strict adherence to net neutrality is not
desirable in certain cases, for example, to support streaming
services and real-time online gaming, and discuss the
implications this may have on resource sharing and network
performance.

21.1 Basic Net Neutrality

Definition 21.1 Net Neutrality
 Net neutrality is the principle that all communication on the Internet
shall be treated equally and without any form of discrimination by
the Internet service providers (ISPs).

With net neutrality in force, there shall be no discrimination of the

transmission of data based on the identity of the sending or the
receiving users, the content of the data, or the associated application.
This means that data packets transmitted on the Internet should be
subject to the best-effort paradigm and handled on a first-come-first-
served basis. This also means that the ISPs cannot perform any kind of
blocking of applications, data rate throttling of specific applications, or
any differentiated treatment of data packets based on the identity of the
sender or the receiver of the data packets. With full net neutrality in
force, even advanced network management to support service quality is
forbidden; for example, schemes that give different treatments to voice
communication compared to other Internet traffic. ► Section 21.2
discusses why 4G and 5G mobile networks are an exception to the strict
rule of net neutrality.
Net neutrality is an important part of the Open Internet Rule, in
which the Internet shall be open and accessible for everybody without
any kind of discrimination. Moreover, under the Open Internet Rule, any
consumer’s access to or usage of the Internet should not be driven by
financial motivations of the ISPs. Net neutrality effectively reduces the
ISPs to a carrier of bits between senders and receivers. Any
involvement of the ISP in higher layer functionalities or other services
is in general not compatible with net neutrality (Murray, 2016).
The term net neutrality was coined by Professor Tim Wu in his
paper Network Neutrality, Broadband Discrimination (2003) (Wu,
2003) and has been the target of much political debate since then. The
main issue is whether or not the principle of net neutrality shall be
enforced on the Internet.
Proponents of net neutrality claim that equal treatment of all
services will foster innovation on the Internet—for example,
development of new apps and services—and safeguard a democratic
platform in which all information is treated equally. If net neutrality is
not enforced, ISPs may offer fast lanes to established and dominating
application service providers (ASPs) for an extra fee, an action which
will strengthen the dominance of certain ASPs. ISPs may also block or
throttle Internet speeds for ASPs competing with the ISP’s own
services, for example, an ISP offering traditional voice communication
in competition with Skype throttles the Internet speed of Skype to gain
competitive advantage.
Net neutrality is also required to ensure that the Internet remains a
democratic platform. This is so because, without net neutrality in force,
ISPs may block content for some reason, for example, political motives
violating free speech and democracy. Net neutrality is thus not only a
technical or economic issue but a central point concerning human
rights and the evolution of democratic and political standards.
Opponents of net neutrality claim that net neutrality will reduce
incentives of the ISPs to invest in the network and thus slow down
further Internet adoption and technological progress and innovation.
The ISPs claim that it will be hard for the ISPs to get sufficient returns
on infrastructure investments if they cannot charge large application
service providers—such as YouTube and Netflix—extra for their
enormous usage of the network. Therefore, it is not surprising that the
main stakeholders that favor net neutrality include ASPs such as
Facebook, Netflix, and Microsoft, while the stakeholders opposing net
neutrality include mostly ISPs.
Many countries have passed legislations on net neutrality. Among
them, Chile was the first country to pass full net neutrality legislation in
2010. As a consequence of this law, zero-rated applications (see ► Sect.
21.5)—including Facebook Zero—are no longer available in Chile. In
the USA, net neutrality has been a source of conflict since the 1990s.
The Federal Communication Commission (FCC) published in 2010 a set
of six net neutrality principles to govern the providers of Internet
access (i.e., ISPs) (Preseving the Open Internet, 2010). After that, net
neutrality in the USA has been subject of political debate several times.
In 2017, President Donald Trump and the FCC reversed the rules, and
the USA (except California) is still (by March 2021) without regulations
concerning net neutrality (Kastrenakes, 2017; Kelly, 2019).
Six principles are termed the FCC Open Internet Order. Though they
may no longer be in force in the USA, they are probably the most
concise definition of net neutrality that exists:
1. Transparency: Consumers and innovators have a right to know the
 basic performance characteristics of their Internet access and how
their network is being managed.  
2.
No blocking: Consumers and innovators have a right to send and  
receive lawful traffic, to go where they want, to say what they want,
to experiment with ideas—commercial and social—and use the
devices of their choice. The rules thus prohibit the blocking of
lawful content, apps, services, and the connection of devices to the
network.
3.
Level playing field: Consumers and innovators have a right to a level  
playing field. No central authority, public or private, should have
the power to pick winners and losers on the Internet.
4.
Network management: Broadband providers need meaningful  
flexibility to manage their networks to deal with congestion,
security, and other issues.
5.
Mobile: The principle of Internet openness applies to mobile  
broadband.
6.
Vigilance: Promptly enforcing the rules to be adopted and vigilance  
in monitoring developments in areas such as mobile and the
market for specialized services, which may affect Internet
openness.
In the EU, net neutrality is laid down by article 3 of EU regulation
2015/2120: safeguarding of open Internet access (Regulation (EU)
2015/2120 of the European Parliament and of the Council, 2015). This
regulation is a part of the union’s Digital Single Market policy and was
announced in 2015. The law broadly ensures net neutrality in the
EU/EEA zones. However, countries within the union may specify
stricter net neutrality rules than those in the EU regulation. This is done
in the Netherlands and in Slovenia. The EU regulation on net neutrality
has been criticized for being vague and open up for prioritization of
“specialized services” such as remote surgery and driverless cars. Such
prioritization is in violation of the principles of net neutrality as
differentiated treatment of data packets in the network is needed to
improve the quality and precision of such services. Another criticism of
the EU regulation is that it opens up for zero-rated applications.
China, on the other hand, has not enforced net neutrality. On the
contrary, China blocks or restricts certain services, for instance,
Facebook, within China for political reasons.
Mobile systems represent a particular problem concerning net
neutrality as explained next.

21.2 5G and Net Neutrality

5G systems may violate some of the net neutrality principles. The 5G
technology is an extension of the Internet to mobile devices. For
technical reasons (e.g., bandwidth efficiency, dynamic bitrate allocation,
and forward error correction to reduce packer-loss rate), the mobile
network provider must know more about the type of signal to set up a
connection between the mobile user and the network. For this purpose,
some degree of signaling identifying the service is required over the
interface and in the mobile network. This is called the IP multimedia
subsystem (IMS). In addition, the provider may offer different priorities
and meet certain latency requirements for different services as shown
in ◘ Table 21.1. Technically, this is referred to as quality of service (QoS)
management. In principle, QoS management makes it possible for the
mobile provider to charge different services differently and to offer
different user experience depending on the type of service. This may
infringe with the net neutrality principle. It is then up to national
regulations and market supervision to ensure that net neutrality is not
violated.
Table 21.1 QoS parameters for 5G mobile networks

Service Resource Priority Packet delay Packet Example services

class type (latency) loss
(ms) ratea
1 GBRb 2 100 10−2 Telephony

2 GBR 4 150 10−3 Person-to-person video (video

telephony)
3 GBR 5 300 10−4 Real-time TV quality video stream
Service Resource Priority Packet delay Packet Example services
class type (latency) loss
(ms) ratea
4 GBR 3 50 10−3 Real-time interactive games

5 Non-GBR 1 100 10−4 IMSc signaling

6 Non-GBR 6 100 10−3 Simple interactive games and live
streaming (video, podcast)
7 Non-GBR 7 300 10−4 Common Internet services (WWW, e-
mail, over-the-top services, etc.) and
SMS

Authors’ compilation, based on data from 3GPP Specification TS 23.501.

System architecture for the 5G system. Section 5.7
aA loss rate of 10−2 means that at most one out of 100 packets is

allowed to be lost on average

bGBR = minimum guaranteed bitrate on the radio interface
cIMS = IP multimedia subsystem (signaling and packet-transfer

management)

3GPP defines four QoS parameters for 5G systems resulting in seven

service classes as shown in the table.
– Minimum guaranteed bit rate. Some services such as telephony, real-
time video, and real-time gaming require a minimum guaranteed bit
rate, while services such as web-browsing services and buffered
video-streaming services require no such guarantee. Telephony and
real-time video require a minimum guaranteed bitrate because
speech samples and video pixels are generated at a fixed rate at the
encoder and the signal must be fed to the decoder at the same speed
for proper decoding.
– Priority has to do with the priority given to packets queuing up at
buffers. A packet with priority 1 will be put in the front of the queue
but after priority 1 packets that are already queued up. Equivalent
rules apply to the other priorities. Generally, real-time services such
as telephony, signaling, and real-time gaming should be handled first,
while packets belonging to a web-browsing service may wait.
– Packet delay has to do with latency. For some services latency is
critical such as for real-time gaming, while for buffered streaming
services and web browsing, latency is not critical. For other services
such as telephony, latency is a problem. The roundtrip delay (i.e., the
time from one party finishes speaking to the earliest time the reply
can be received from the opposite party) should preferably be less
than 300 milliseconds. For communication over geostationary
satellites, the roundtrip delay is about 500 milliseconds. At this delay,
it starts getting difficult to maintain a fluent conversation without
interruptions. For two satellite hops in tandem, the roundtrip delay
is 1 second. In that case, fluent conversation is difficult.
– Packet loss should usually be kept low for data services. On the other
hand, voice and video services may tolerate much higher packet loss
probability than, for example, downloading of a web page. This is so
because a lost speech sample or video pixel may be replaced by the
previous sample or pixel without being noticed by the listener or
viewer. A lost packet containing part of a web file may make the file
unreadable.
Note that it is not possible to distinguish between traditional
Internet services (web search, e-mail, over-the-top services (e.g.,
Skype), Facebook, Twitter, YouTube, and so on) since all of them
belongs to the same service class (class 7), so that the original idea of
net neutrality is not violated by 5G. 5G illustrates that full net neutrality
may not always be possible for technical reasons. In general, high-
quality VoIP, real-time video streaming, and real-time interactive games
offered over fixed Internet interfaces will also require QoS management
similar to 5G. The net neutrality regulations in the USA and EU open up
for this possibility.

21.3 Device and Search Neutrality

Two terms related to net neutrality are device neutrality and search
neutrality.

Definition 21.2 Device Neutrality

 Device neutrality means that any application should be able to run on
any device and that it is possible to connect the device to the
network of any ISP without differentiation of price or quality.

The standardization of the technical infrastructure of ICT supports

these requirements.

Definition 21.3 Search Neutrality

Search neutrality means that search engines shall return unbiased
results to the user and be optimized to provide the most relevant
results based solely on the search keywords provided by the user.

Hence, commercial interests such as promotion of paid services or

services owned by the company offering the search engine should not
be a parameter in the algorithm calculating the results of the search.
Device and search neutrality have a less—if any—legal basis compared
to net neutrality. However, there have been legal cases where
companies—most notably Google—have been fined for breaching
search neutrality.
Google was accused of favoring services from their own ecosystem.
For this practice, Google was fined €2.42 billion in 2017 by the
European Commission (Vincent, 2017). More specifically, Google was
fined for manipulating the search results in Google Search to favor
results from Google Shopping—a Google service that allows users to
search for products on e-commerce websites. Competing price
comparison services were—according to the judgment—intentionally
put far down on the list of Google Search results in such a way that
consumers often ignored these results. This is a clear violation of
search neutrality since Google used its monopoly dominance in the
search market to favor its own products (Google has over 90% market
share in the search market in Europe). Google has appealed the
decision. Andrew Odlyzko predicts that device and search neutrality
may become “hot topics” in the future representing “the next step” in
regulating the Internet (Odlyzko, 2009). The legal case of EU vs. Google
Shopping may mark the start of recognizing the importance of search
neutrality.
 To make the discussion of search neutrality less clear, it should be
noted that search engines are standard devices in newspaper
databases, product catalogs, address lists, and social media. For some of
these, research neutrality is irrelevant but not for all of them. Both
Amazon and Facebook are accused for strongly biased search (Shavins,
2014). Publishers may pay Amazon to get a higher rank in their
bestselling lists and to include their books among search results.
Facebook is accused to filter news to match user “preferences” based on
previous behavior of the users.

21.4 Business Implications of Net Neutrality

The business implications of net neutrality are significant. With net
neutrality in force, ISPs cannot discriminate data from over-the-top
(OTT) providers—e.g., Netflix, Skype, and WhatsApp—to curb
competition with their own equivalent services. Net neutrality works as
a barrier and strengthen the division between the business domains of
the ISP and the ASP. The ISP has less—if any—opportunities to enter
the business domain of the ASP if net neutrality is enforced. For reasons
explained in ► Chap. 4, full net neutrality divides the business domains
of the ISP and the ASP in such a way that the ISP becomes the
transporter of bits—a commodity—and the ASP becomes the provider
of the services that uses these bits in its service design. This means that
the ASP builds its business on the bit-transportation capabilities
provided by the ISP and no other features of the network. Therefore,
net neutrality has consequences for the business of the ISP since the
ISP is reduced to a commodity and cannot enter the—sometimes
lucrative—business domain of the ASP. This is one of the reasons why
many major ISPs oppose net neutrality. The 5G technology and other
broadband Internet access technologies may alter this picture as
described in ► Sect. 21.2.
The ASP may sometimes be willing to pay the ISP extra for caching
parts of their content material closer to the consumer. This may be
regarded as advanced network management to increase the quality of
the ASP service. In this case, certain types of traffic are given priority
over other types of traffic but without discriminating traffic belonging
to the same type of service. This can be seen as a “mild” violation of net
neutrality. 5G mobile systems are planned to exploit these capabilities
to reduce latency and traffic load in Internet of Things applications by
providing storage and processing capabilities at the radio interface
(edge computing) (Shaw, 2019).
The Internet was originally designed as a “dumb pipe” or “dumb
network” only capable of forwarding IP packets. David Isenberg
denoted the Internet the “stupid network” interconnecting intelligent
terminals in contrast to the “intelligent network” interconnecting
stupid terminals in the telephone network (Isenberg, 1997). There is no
“intelligence” or functionalities built into the IP network apart from
those required for routing IP packets hop-by-hop from the sender to
the receiver. Management and control functionalities are limited to
keeping updated routing tables in the IP routers. If required by the user
application, reliable end-to-end communication is ensured by the end-
to-end protocol Transmission Control Protocol (TCP); if end-to-end
reliability is not required, the simpler User Datagram Protocol (UDP) is
used. These protocols contain an address called port number
identifying the software the receiver must instantiate in order to
interpret the information content of the packet. Sometimes the port
number is unique for a certain service. However, in most cases,
knowing the port number is not enough information to distinguish
between various services. To do so, deeper analysis of the content of the
packets is required. Differentiated treatment of Internet traffic will
require some form of deep packet inspection; see ► Box 21.1.

Box 21.1 Deep Packet Inspection

Deep packet inspection (DPI) implies, as a minimum, that the
network provider reads the UDP/TCP headers to obtain the port
numbers to identify the type of protocol being used (shallow DPI)
and, if possible, analyzes the actual information in the packet itself
(proper DPI).
Sometimes shallow DPI is enough for special treatment of the
packet; in other cases, for example, distinguishing between different
services using the World Wide Web, deeper analysis is required.
Information obtained in this way, together with the addresses of the
sender and the receiver, is used to differentiate the traffic; that is,
decide how the packets shall be treated in the router queues (e.g., be
 given priority, throttled, or blocked). Proper DPI is not possible for
applications using Transport Layer Security (TLS) such as https since
information beyond the port numbers is encrypted and cannot be
decoded by the inspector of the packet unless the encryption key has
been compromised or provided voluntarily or by law to the authority
inspecting the packets. DPI is also impossible for packets sent over
virtual private networks (VPN) because these networks are usually
protected using the strong IPsec encryption protocol and tunneling
techniques where even the addresses of the sender and receiver are
hidden so that there is no information available for discriminating
the traffic.
DPI is used within local networks both at the sending and
receiving end (e.g., in stateful firewalls and for e-mail filtering) to
detect illegal operations, intrusion attempts, spam, and malware and
to prevent sending of protected information into the open Internet.

21.5 Zero-Rating
Some ISPs, in collaboration with selected ASPs—such as Wikipedia and
Facebook—offer zero-rated access to the Internet. This means that
consumers get free Internet access, but then only for accessing selected
applications or services. Put in another way, unlimited data volumes are
provided for a specific application to users opting for zero-rating
access. This practice is in conflict with the current strict definitions of
net neutrality since it differentiates Internet access based on
application—one service can be accessed for free, while another
competing service requires paid access or are not accessible at all.
One example of zero-rating is Wikipedia Zero offering free access to
Wikipedia on mobile devices in some countries in collaboration with
selected ISPs (Russell, 2013). The program was launched in 2012 and
provided free access to over 800 million people, mostly in developing
countries. After receiving criticism for net neutrality violation,
Wikipedia Zero discontinued the program in 2018. In some of the areas
where Wikipedia Zero was deployed, it was, in fact, the only choice for
many people to access the Internet. In these countries, Wikipedia Zero
became synonymous to the Internet. In lack of popular services such as
YouTube—which was only available to those with a regular mobile data
subscription—copyrighted material started to be spread via Wikipedia.
This material was mostly removed by Wikipedia editors; however, it
also meant that these editors collectively became a central force in
deciding what should be available on the Internet through Wikipedia
Zero.
Another example of zero-rating is Facebook Zero, a program
providing free access to Facebook. Launched in 2010, it currently
provides free access to Facebook in collaboration with selected ISPs in
more than 30 countries, both developed and developing countries. In
Nigeria, Indonesia, India, and Brazil, where the Facebook Zero program
is available, more than 50% of the people believe “Facebook is the
Internet” (Mirani, 2015). Compared to Wikipedia Zero, Facebook Zero
is more questionable from a net neutrality viewpoint. This is because
Facebook is a commercial service and not a nonprofit service like
Wikipedia. Providing free access to Facebook changes the competition
in the social media market and may further increase Facebook’s
dominating position in this market.
Twitter has also initiated a zero-rating program—Twitter Zero—
which is available for subscribers of selected ISPs in more than six
countries.
Zero-rating gives the ISPs the power to select winners in the digital
markets motivated by how much they are willing to pay for zero-rating
access of their service. Even though zero-rating means free services for
the users, the cost of providing this service is in many cases paid by the
ASP. Consumers, when everything else is equal, prefer services that
have zero-rated access compared to paid access. Therefore, starting a
zero-rating program for a service may be a way to circumvent
competition, thereby creating a virtual monopoly for this service.
One issue concerning zero-rated content is that ASPs may offer
access to their websites or services for free also in cases where these
services are not the best services for the consumers. For instance, a
bank with high interest rates for loans may pay an ISP to offer free
access to its website to attract customers. This will have an undesirable
effect on the free market for loans. A particularly vulnerable target
group for such practices is poor people with few other opportunities to
access the Internet than through a zero-rated service.
21.6 Conclusions
Net neutrality has transformed the Internet into a formidable arena for
innovation of new services and applications. The technology has also
created entirely new business concepts. This includes concepts such as
sharing economies, social media, e-commerce, streaming of music and
films, multiplayer online games, distance learning, telemedicine, and
much more. The main reasons have been that:
– The extremely simple Internet technology itself does not
discriminate between the different content the data packets may
contain—the Internet itself is completely neutral.
– The evolution of the information and communication technology is
characterized by steady progression toward higher data rates, wider
range of mobile applications, more storage capacity, and increased
processing speed. Hence, the arena for experimentation and
innovation is expanding, steadily allowing new opportunities to be
explored.
– The Internet protocol is such that the ISPs, in most cases, cannot
identify the type of content the data packets contain and for what
purpose they are sent. In democracies, regulations also oblige them
to treat all users equally. The same applies to services except in cases
where differentiation is necessary for technical reasons, e.g., real-
time streaming versus downloading of webpages. This also leads to
flat price structures independent of application and, to a large extent,
also to the volume of data exchanged.
Net neutrality obviously restricts the business opportunities of the
ISP. Therefore, the strongest opponents against net neutrality are the
ISPs. They claim, among others, that net neutrality reduces the
willingness to invest in advanced fiber-optic technologies and
broadband networks. Proponents argue that this is, by fare,
counterweighted by innovations in new technologies and applications.
The proponents include application service providers, content
providers, research communities, and consumer organizations.

Questions
1. What are the effects of strict net neutrality on picture and sound
 reproduction on video conferencing?
 
2.
Why is it possible for operators of 5G mobile network to levy  
different charges for the telephone service and data
communication?
3.
Does the Apple ecosystem—iPhone and App Store—constitute a  
violation of device neutrality?

Answers
1.
With strict net neutrality, packets are handled by network  
routers using the first-in-first-served principle. This means that
a packet received at the router is queued until all earlier packets
in the queue have been forwarded. Moreover, packets belonging
to the same service may be forwarded on different routes with
different propagation delay. This results in unpredictable
variation in the time adjacent packets are received by the
decoder. Real-time video, music, and voice are extremely
sensitive to this type of jitter that, in the worst case, may distort
the picture or garble the speech.
2.
When accessing the network, the mobile terminal indicates the  
type of services, e.g., telephone call, data call, or one of several
other categories. This information is required by the network to
allocate bandwidth (e.g., narrowband for telephony, wideband
for video streaming) and to indicate how the call shall be
handled by the network (real-time priority, minimum required
bandwidth, no restriction). The mobile ISP then knows the
category of service the user wishes and may, therefore, levy
different charges for different service categories. This is
violation of net neutrality necessary for technical reasons (for
efficient usage of frequency recourses and for guaranteeing a
minimum quality of service).
3. Yes. This is because apps that run on the iPhone can only be
downloaded from the App Store. Apps not available on the App  
 Store cannot be run on the iPhone. Furthermore, apps on the
App Store cannot be downloaded on devices other than the
iPhone.

References
Isenberg, D. (1997, August). Rise of the Stupid Network: Why the Intelligent Network was once
a good idea, but isn’t anymore. One telephone company nerd’s odd perspective on the changing
value proposition. Computer Technology.

Kastrenakes, J. (2017, December 14). The FCC just killed net neutrality. The Verge.

Kelly, M. (2019, May 6). Democrats push new bill to write net neutrality into law, but can it
pass? The Verge.

Mirani, L. (2015, February 9). Millions of Facebook users have no idea they’re using the
Internet. Quartz.

Murray, A. (2016). Information technology law: The law and society. Oxford University Press.

Odlyzko, A. (2009). Network neutrality, search neutrality, and the never-ending conflict
between efficiency and fairness in markets. Review of Network Economics, 8(1), 40–60.

Preseving the Open Internet. (2010, December 21). Broadband industry practises. Report and
order. Federal Communication Commission.

Regulation (EU) 2015/2120 of the European Parliament and of the Council. November 25,
2015.

Russell, B. (2013, February 21). Wikipedia zero wants to bring Wikipedia to mobile users
without a data plan. TechnoBuffalo.

Shavins, N. (2014, July 2). Are Google and Amazon the next treat to net neutrality? Forbes.

Shaw, K. (2019, November 13). What is edge computing and why it matters. Network World.

Vincent, J. (2017, June 27). Google fined a record €2.4 billion by the EU for manipulating search
results. The Verge.

Wu, T. (2003). Network neutrality, broadband discrimination. Journal of Telecommunications

and High Technology Law, 2, 141–179.
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
H. Øverby, J. A. Audestad, Introduction to Digital Economics, Classroom Companion: Business
https://doi.org/10.1007/978-3-030-78237-5_22

22. Digital Regulation
Harald Øverby1   and Jan Arild Audestad1  
(1) Norwegian University of Science and Technology, Gjøvik, Norway

 
  Harald Øverby (Corresponding author)
Email: haraldov@ntnu.no

  Jan Arild Audestad
Email: jan.audestad@ntnu.no

Keywords Market regulation – Cyberlibertarian – Cyber-paternalist –

Pathetic  dot

Learning Objectives
After completing this chapter, you should be able to:
– Explain why and how mobile communication is regulated.
– Discuss the complexity of regulating the Internet not only
because of its inherent complexity but also because there is
an ongoing conflict between opponents and proponents of net
neutrality and the need to regulate the network.
– Discuss why legislation is not the only way to regulate the
market but that the market itself, the technology, and the
public also contribute to such regulation.

22.1 Introduction
Since the early 1980s, competition has gradually been introduced in the
telecommunications market, first in the UK in 1982 for regulating the
market for cellular mobile communications. In the rest of Europe,
mobile communications were opened for competition in 1992. In 1998,
all telecommunications in Europe were de-monopolized and opened for
general competition. This development and the extensive use of the
Internet have generated a demand for regulating the
telecommunications market to make it a level playing field.
Market regulation for ICT can be defined as follows. (Note that this
definition is not essentially different from the regulation of other
markets.)

Definition
Regulation of the ICT market is the intervention of governmental,
legal, social, economic, or technological authorities, by rules or
procedures, to restrict the freedom of operations for market
participants (in particular, mobile operators, Internet service
providers, and application service providers) and to target the
evolution of the market.

There are several motives for regulation:

– To avoid market failure such as formation of monopolies
– To foster fair competition
– To secure that the users have correct and adequate information
about the market
– To ensure affordable access to the ICT infrastructure, thereby
satisfying collective needs of the public
– To protect individuals against unethical business conduct and abuse
of personal data
– To promote professional and ethical conduct of market participants
– To stimulate peer-based service innovation and development of new
technologies
The regulations apply to fixed and mobile network operators, user
access providers, Internet service providers, application service
provides, and content providers. This chapter examines more deeply
the regulation of mobile communications and the Internet. These are
regulations by law. ► Section 22.4 reviews how social norms, the
marketplace itself, and technology also contribute to regulate the
market.
22.2 Mobile Network Regulations
The most comprehensive overview of regulations of the ICT market is
the Telecommunications Regulation Handbook (Black & Srivastava,
2011). This section and the next are based on this source and rules
applied by the Norwegian Communications Authority (Nkom) (Nkom,
2017).
The main motive for regulating the mobile market is to build the
foundation for a competitive market—or in other words, to avoid
market dominance by one operator and to stimulate new operators to
enter the market. To achieve this, a large number of market attributes
have to be regulated and monitored.

22.2.1 Fair Competition

When Europe opened for full competition of mobile communication in
1992, one of the first company in each European country to established
itself as mobile network operator (MNO) was the operator owning the
entire telecommunications infrastructure of that country—the
government-owned de facto monopoly (also called the incumbent). The
incumbent had thus an enormous initial market power. To reduce the
market power of the incumbent as MNO, the authorities required the
MNO to commercially separate from the other business areas of the
incumbent and that the conditions for interconnecting to the fixed
network and for the use of infrastructure components (e.g., to connect
base stations and exchanges) owned by the incumbent to be the same
for all MNOs, including its own MNO.
The MNO must have access to exclusive slots in the radio-frequency
spectrum. The amount of spectrum allocated for mobile
communications is scarce, and there is room for only a few MNOs in the
same region. Fair competition for frequency resources is achieved by
dividing the spectrum into slots and then auctioning each slot to
existing or new operators. This allowed only a few operators in each
country. To increase competition, the market was also opened up for
resellers and mobile virtual network operators (MVNOs). Resellers buy
bulk traffic capacity from MNOs and resell it under their own brand to
the customers. MVNOs own some network infrastructure but buy
access to the wireless infrastructure from MNOs. Resellers and MVNOs
need no frequency licenses to operate. The regulation authority
enforces competition rules and supervises that resellers and MVNOs
meet fair competition in the mobile market. See also ► Chap. 5 for
more about resellers and MVNOs.
The mobile market is an oligopoly with just a few MNOs in each
country. Some of these MNOs may have market power big enough to
take actions that alter competition or establish new market rules. These
are referred to as dominating MNOs. There may be several dominating
MNOs in a country with approximately equal market shares. The main
objective of market regulation is then to hinder that dominating MNOs
can misuse their market power to drive competitors out of the market,
hinder new entrants to enter the market, or unduly exploit the
customers by overcharging. The following is a list of competition
problems that may arise in the mobile market and must be mitigated by
market regulations.

22.2.2 Denial of Interconnection

MNOs are value networks (see ► Chap. 8 for definition) that benefit
from interconnecting with other national or international MNOs and
fixed networks to make their network of relationships between users as
big as possible. Full interconnectivity in the international telephone
network is governed by rules set up by the International
Telecommunications Union and universally endorsed by the member
countries. These requirements apply to both fixed and mobile
telephone networks.
However, an MNO with dominating market power may squeeze new
entrants out of the market by denying them interconnection or call
termination. This means that users of the new entrant cannot call users
of the MNO and, thereby, reducing the value of the new entrant
dramatically. This conduct is also referred to as denial of traffic
termination. One of the responsibilities of the regulator is to supervise
that such actions do not take place.

22.2.3 Excessive Pricing

The terminating MNO is in a monopoly situation since this is the only
network in which a particular call can end up (i.e., where the called user
lives or are temporarily located). This allows the terminating network
to decide the price for connecting the called user, a price the calling
network (and the user) must accept. If the price claim is not accepted,
the call is rejected by the terminating MNO. The terminating MNO may
then be tempted to levy excessive charges. To avoid such behavior, the
regulator may set a price cap for call termination, making the prices
more predictable for the user. However, lower bilateral termination
prices may be negotiated between MNOs to support roaming users.
In EEA, excessive pricing is avoided by the price cap method; that is,
the termination price of all MNOs in the EEA region must be equal to or
lower than the price cap set by the national regulator. Outside the EEA,
there are several countries in which the termination price is not
regulated and can be set independently by the termination MNO.

22.2.4 Cross-Subsidizing
Cross-subsidizing means to charge excessive prices for one service (the
subsidizing service) and to use the additional earnings to reduce the
charges for another service (the subsidized service). The major source
for cross-subsidizing in the telecommunications market is high
termination charges. These earnings may be used to subsidize another
service and thereby obtain competitive advantages for that service.
Cross-subsidizing may, to a large extent, be avoided by price-cap
regulation of call termination charges as explained above. Cross-
subsidizing between fixed and mobile network operation is avoided by
requiring that the subsidiaries offering fixed and mobile services are
commercially separated.

22.2.5 Price Discrimination

The terminating MNO may charge lower termination charges for calls
from MNOs belonging to the same group (e.g., a subsidiary in another
country) and from other MNOs with which the terminating MNO has
particular agreements (e.g., bilateral roaming agreements). Such
practice may upset competition and should be avoided by regulations.
Price discrimination may also be used for cross-subsidizing by
charging low termination charges from own subsidiaries and high
charges from other MNOs.

22.2.6 Lock-In of Customers

Customers may be locked in by contractually binding the customer for a
period of time and to enforce economic penalties if the customer leaves
the provider before the end of the contractual period. This may be done
by offering cheap mobile phones to customers who accepts the contract
and mobile phones for market price for those who do not. Another
method is SIM lock where the mobile phone will not accept a SIM from
a competing MNO until the lock has been removed or after a certain
time. In some countries, these activities are illegal, while they may be
allowed to a certain extent in other countries.

22.2.7 Non-Price Discrimination

There are also several factors other than price that may twist
competition in an undesirable direction. Examples are:
– Dragging out interconnection negotiations, thereby slowing down
the market growth of the competitor
– Deliver insufficient interconnection specifications, also slowing down
competition or making interconnection more expensive for the
competitor
– Deliver stripped down functionality, thereby disallowing the
competitor access to some interconnection services
– Reduced quality of technical interfaces (e.g., throttled data rate, slow
connection establishment, long latency, and so on)
– Unwarranted requirements (e.g., liabilities in case of network
failures)
Negotiating the interface between MVNOs and MNOs is particularly
complicated because it includes both commercial and technical aspects
that are much knottier than the interconnection of ordinary MNOs
(Audestad & Gaivoronski, 2001).

22.3 Internet Regulations

Internet played no role in the de-monopolization of
telecommunications. During the 1980s, the Internet grew in size to
interconnect universities, laboratories, and innovative industries for
exchange of emails, ideas, documents, and other information. The
Internet was more or less a research network not recognized by
telecommunications operators, except that they provided backbone
connections between routers at universities and laboratories, thereby
building up a global data network without central governance. At the
same time, the telecommunications monopolies developed their own
standards for data communications: an inflexible and short-lived circuit
switched data network based on the same technology as the telephone
network and a packet switched data network (the X.25 network) less
flexible and more expensive to build than the Internet. The
telecommunications operators levied charges for the use of these
networks (e.g., a fixed price per data packet sent) to regain investments,
cover operational costs, and increase their revenues. The use of these
networks never gained momentum since the need for data
communications outside the research community was almost absent at
that time and the Internet offered the type and volume of data
communications academia needed free of charge.
In the mid-1990s, the general public discovered that the World
Wide Web offered them opportunities that had never existed before, for
example, access to electronic newspapers, public information, and
entertainment. Moreover, people could access this information from
anywhere and at any time, thereby offering them a new freedom of
choice. The World Wide Web was designed for implementation on the
Internet, and in order to meet the new demands from the users, the
telecommunications operators also started building their own
fragments of the Internet to offer web browsing services. This replaced
the need for other data networks, and the Internet soon became the
only data network interconnecting users worldwide. As a consequence,
the telecommunications operators could no longer levy charges for use
of the Internet in the same way as in the telephone network.
When the 3G mobile network was introduced in 2000, the
application of Internet services on mobile phones made a substantial
jump. International telecommunications agreements included the
telephone network but not the Internet. This had one particular effect
on prices while roaming to another country. While the roaming prices
for telephony where regulated and agreed upon by the mobile
operators, they could levy any price for data calls to and from users
roaming from another country. In several countries, the price roaming
users had to pay for data calls was excessive. Moreover, some mobile
networks generated data traffic to roaming mobile terminals such as
unnecessary updating and messages. The user had no control over this
traffic but had, nevertheless, to pay for it. For this reason, the mobile
phone allows the user to disconnect data calls while roaming into
another country while maintain roaming for telephone calls and SMS.
In 2017, the EEA forbid member states to continue this practice: data
calls should be handled in the same way as telephone calls and SMS
while roaming in the EEA. The home network of the mobile user should
inform the user that the mobile had roamed into a safe network and
warn the user if the phone had roamed to a network where data
roaming prices may not be regulated.
The increased popularity of the Internet in the 1990s triggered the
rise of the cyberlibertarian movement (Borsook, 2001). The
cyberlibertarian’s main opinion is that the Internet should not be
regulated by international, regional, or national laws. They claim that
the Internet—or cyberspace—does not follow national borders. Data
packets are often routed over several countries and legal jurisdictions
from the sender to the receiver. Data from a single transaction could
even take different paths in the network crossing different national
borders. The legislation of a single nation can, therefore, not be applied
to the Internet. The Internet user, including ASPs and content
providers, could then exploit regulation arbitrage, meaning that the
laws of the country with the most liberal laws and regulations would be
used, for example, by placing the servers supporting the service in low-
tax countries.
The cyberlibertarians argue that the Internet should be allowed to
govern itself, democratically, and without any central control.
As a response to the cyberlibertarian movement the cyber-
paternalists came on to the scene. They claimed—contrary to the
cyberlibertarians—that the Internet should indeed be regulated to
function properly. Even though cyberspace invisibly crosses national
borders, cyberspace is built up of equipment—routers, switches,
terminals, mobile stations, fiber-optic cables—owned and used by
people or companies under the jurisdiction of the legal framework of a
country. The question raised by the cyber-paternalists is not whether
cyberspace should be regulated or not, but rather whether such
regulations could be done by applying existing laws or by developing
new laws and rules particularly for the cyberspace.
Today, most academics and decisions makers agree that the Internet
both can and should be regulated. Indeed, legal frameworks of many
countries have been or are about to be updated because of the
widespread use of the Internet and other related information
technologies. One major reason for regulating the Internet is to prevent
market dominance. Because of strong network effects and that the
marginal cost associated with many digital goods is zero, several
markets in the digital economy will be dominated by de facto
monopolies if regulations are absent.
One example of a de facto monopoly is Facebook. The market of
Facebook is not regulated and, therefore, prone to market failure.
Dominating network effects and path dependence have turned
Facebook into a de facto monopoly as we have deliberated several times
in this book (► Chaps. 9, 11, and 13). Moreover, the product Facebook
offers the users has zero marginal cost and is provided to the users for
free. The question is whether markets subject to such conditions can be
regulated at all. One concern is that if Facebook is split into two
competing companies (antitrust regulation), the market evolution will
most likely follow a path where one of them ends up as a new
monopoly, while the other will disappear from the market. In ► Chap.
11, a simple mathematical model shows that this evolution is indeed
possible and quite likely.
Other areas of regulation on the Internet are to protect consumer
privacy, to ensure that telecom operators pay for their use of public
goods (e.g., frequencies), and to stop piracy and illegal distribution of
content on the Internet. In general, an increasing number of regulations
of the Internet have been put in force during the past decade. These
include:
– The General Data Privacy Regulation of EU (General Data Privacy
Regulation, 2018).
– Net neutrality as described in ► Chap. 21 is by far the most
important regulation on the Internet limiting the power Internet
service providers have on the evolution of the applications and
service provisions of the network.
22.4 Lessig’s Four Modalities
Regulating the Internet may not only be done by laws and legal
frameworks. The pathetic dot theory developed by Lawrence Lessig
(Lessig, 1997) defines four modalities of regulation, as illustrated in ◘
Fig. 22.1:
– Legal: How the legal framework in a jurisdiction is used to regulate
– Market: How trade, markets, and economic factors are used to
regulate
– Technology: How the technology is used to regulate
– Society: How norms and societal factors are used to regulate

Fig. 22.1 Lessig’s four modalities of regulation. (Authors’ own figure)

Regulating the digital economy—or a specific sector, domain, or

market in the digital economy—can be achieved by using a
combination of the four modalities. The principle is shown in ►
Example 22.1 using the music industry as example.

Example 22.1 The Four Modalities Applied to the Music

Industry
The regulation of music piracy is used as an example. One key
problem in this industry is the violation of copyright and illegal
downloading and spreading of music on the Internet. Such actions
were made possible by the ubiquitous use of the Internet combined
with applications or websites such as Napster and ► MP3.​com. This
was a major issue in the 2000s and still is, however, with less intensity
today since a combination of the modalities described above has been
employed to regulate the issue.
Regulation by Legal Measures: In many countries, downloading
copyrighted material is illegal by law. People downloading and
sharing such material may be prosecuted and punished according to
the laws in their jurisdiction. This is an example of legal measures in
Lessig’s model to regulate software piracy.
Regulation by Market: In the 2000s, new services offering access
to copyrighted media were launched—Spotify, iTunes, and Tidal.
These services created a market for legal access to music and
contributed to regulating the market. This is an example of regulation
by market.
Regulation by Technology: The 2000s also saw the emergence of
technological copyright protection of music and other media by which
copying a specific CD or DVD was not possible. This is the use of
technological measures to regulate piracy.
Regulation by Society: The last of Lessig’s modalities—society—
is about societal actions to regulate piracy. In spite of the laws passed
to regulate piracy, people still in huge numbers continued to
download and spread copyrighted material illegally. This was because
the general opinion of the public was not to view free music
downloading as a crime that should be punished. Campaigns
comparing stealing music and stealing cars as the same thing did not
have any lasting effect on the public. In many people’s opinion, digital
goods are different from physical goods since digital goods are non-
rival, while physical goods are rival by nature: stealing a non-rival
 good is not the same as stealing a physical thing from someone. In this
case, public opinion is a weak regulating force.

The main point is that regulating the digital economy can be achieved
not only by law but also by markets, business models, economic
incentives, technology, design, and societal campaigns. These forces—
or modalities as Lessig termed them—work together and influence one
another. How well a specific service or part of the digital economy is
regulated is the sum of all these effects and their interactions.

22.5 Conclusions
Several sectors of the digital economy must be regulated to avoid
market failure and to create a level playing field for all providers of
technology, services, applications, and content. Regulations also protect
the users against access discrimination and protect them against
excessive pricing and misuse of information about the users and their
preferences, habits, and other personality traits. Some of these sectors
may be difficult to regulate, for example, protection against formation
of monopolies in the application and content provider sector and the
use or misuse of personal information for commercial purposes.
Two of the most important sectors that need regulation are mobile
communications and the Internet. The fixed network is also regulated,
but this regulation is less and less important as telecommunications
now converges rapidly toward a mobile Internet as explained in ►
Chap. 3. These are regulations by law that are governed and supervised
by public authorities. In addition, the market may also be regulated by
the market itself, by means of technology, and by public opinion and
ethics (the pathetic dot theory).
It is particularly difficult to regulate the Internet. The reason is, as
explained in ► Chap. 4, that the Internet is divided into two
commercially independent domains: Internet service providers in
charge of transporting bits and application and content providers
creating, storing, and disseminating information and services.
One important field of regulation is net neutrality, shaping the
Internet into an open and unrestricted laboratory for innovation and
exploitation of new ideas. In some countries, there are strong forces
working against net neutrality and for an Internet where the Internet
service providers alone determine the conditions for using the Internet.
This development may hamper the evolution that have, during less than
two decades, created several millions of new applications on the
Internet.

Questions
1.
Is there a link between human rights and net neutrality?  
2.
What are the incentives for self-regulation in the mobile market  
based on Lessig’s theory?

Answers
1.
Net neutrality is not stated as an explicit human right but is  
regarded as a must for the implementation of rights such
“freedom of speech.” The weaker requirement “right to Internet
access” is a non-binding resolution of the United Nations Human
Rights Council.
2. Incentives:
(a)
 
Law: penalties for not obeying national regulation  
legislation.
(b)
Market: the most important motive is to make the total  
market pie as large as possible by accurately designing the
system in accordance with a common standard. This also
maximizes the market for each operator both with regard
to own customers and visiting customers (roaming). The
marketplace then becomes a common in which providers
can compete on price and customer care to attract and
keep customers.
(c) Technology: the cost of development of the technology is
reduced for each operator if they collaborate to develop a  
common standard. If there are more than one incompatible
t d d th t d d i lik l d i th l
 standard, then a standards war is likely, and, in the long
run, one of the standards will win the war. Building
backward compatibility into the technology both in the
mobile phones and the network ensures a smooth
evolution of the technology.
(d)
Society: Word of mouth may increase the business of the  
MNO that offer the best and cheapest service.

References
Audestad, J. A., & Gaivoronski, A. (2001). Option pricing of mobile virtual network operators.
Telektronikk, No. 4.

Black, C., & Srivastava, L. (Eds.). (2011). Telecommunications regulation handbook (Tenth
Anniversary ed.). The International Bank for Reconstruction and Development / The World
Bank, InfoDev, and The International Telecommunication Union.

Borsook, P. (2001). Cyberselfish: Revers, guilders, cyberpunks, and other silicon valley
lifeforms. Yale Journal of Law and Technology, 3.

Decisions on designation of providers with significant market power and imposition of specific
obligations in the markets for wholesale voice call termination on individual mobile networks
(Market 2). Norwegian Communications Authorities (Nkom). 2017.

General Data Privacy Regulation. (2018). Intersoft consulting.

Lessig, L. (1997). Code and other laws of cyberspace. Basic Books.

Index
A
Access economy
Acquisition
Ad-based free
Adobe PDF
Advertisement
Age of Conan
Airbnb
Albert, Réka
Alibaba
Alipay
Alphabet
Amazon
Anarchy Online
Anderson, Chris
Android
Anything-as-a-Service
Apple
App Store
Apple II
iTunes
Pages
Pay
Application Service Provider
Apps
ARPANET
Assets
Attention economy
B
Baidu
Barabási, Albert-László
Bardeen, John
Bass diffusion model
Bass, Frank
BBC
Benefactors
Betamax
Big Data
Bing
Bitcoin
Blizzard
Blockchain
Bluetooth
Brattain, Walter
Broadcast network
Bundling
Business model
Business Model Canvas
Business-to-Business (B2B)
Business-to-Consumer (B2C)
C
CAESAR
Cambridge Analytica
Churning
Cloud
computing
storage
Commodities
Commodore PET
Common-pool resources
Commons-based peer production
Compact disc (CD)
Complement
Complexity economics
Consumer adoption
Consumer-to-business (C2B)
Content provider
Convergence
Cooperation
horizontal
 vertical
Coopetition
Core competencies
Cost structure
Crowdsourcing
Cryptocurrencies
Bitcoin
Ethereum
IOTA
Litecoin
Ripple
Customer relationships
Customer segments
Consumer-to-Business (C2B)
Consumer-to-Consumer (C2C)
D
Data
analog
digital
Dell
Device provider (DP)
Differential equations
Digital
economy
economy ecosystem
goods and services
market
photography
Digital audio broadcasting (DAB)
Digitization
Dropbox
Dunbar, Robin
DVD
DVORAK
E
E-banking
eBay
E-commerce
EDGE
E-mail
Erdös-Rényi (ER) graph
Ethereum
European Union
Experience economy
Exponential growth
F
Facebook
Firefox
5G
4G
Freemium
Funcom
G
Global system for mobile communications (GSM)
Goods
club
non-excludable
non-rival
private
public
rival
search
Google
Chrome
Docs
Drive
Gmail
Google+
Hangouts
News
Pixel
Play
 Voice
GPRS
Grameen Bank
Grand Theft Auto V
Groupon
H
HBO
Hypertext Markup Language (HTML)
Hypertext Transfer Protocol (HTTP)
I
Imitators
Inflexion point
Information and communication technology (ICT)
infrastructure
Information economy
Infrastructure provider
In-house production
Innovation
disruptive
Innovators
Instagram
Integration, vertical
Intel
Internet
banking
economy
of Things
protocol
routers
Internet Engineering Task Force (IETF)
Internet Service Provider
IOTA
iPhone
iPod
K
Key activities
Key partners
Key resources
Kickstarter
Klarna
Kodak
L
Latency time
Layered Internet model
Layering
Lilienfeld, Julius Edgar
Linear growth
LinkedIn
Linux
Litecoin
Lock-in
Long tail
M
Machine learning
Marginal cost
Market adoption
Market capitalization
Massive multiplayer online game
Massive open online course
MasterCard
Merger
Mergers & acquisitions (M&A)
Messenger
Metcalfe, Robert
Microsoft
Internet Explorer
OneNote
Outlook
PowerPoint
Word
Microsoft Disk Operating System (MS-DOS)
Minecraft
Monopoly
Mosaic
Multimedia Messaging Service
Multi-sided platform
Myspace
N
Nakamoto, Satoshi
Negative feedback
Net neutrality
Netflix
Netscape
Network effects
cross-side
direct
indirect
negative
positive
same-side
Network laws
New York Times
Nobel Prize
Nordic Mobile Telephone
NSFNET
O
Odlyzko, Andrew
Online banking
Open source software
Optical fibers
Organic growth
Osterwalder, Alexander
Over-the-top services
P
Packet switching
Panasonic
Path dependence
PayPal
Personal computer (PC)
Popcorn Time
Porter, Michael
Positive feedback
Postmates
Prosumer
Q
QWERTY
R
Recommendation systems
Reddit
Riemann zeta function
Ripple
S
Safari
Samsung Galaxy
Sarnoff, David
Satellite networks
Saturation
S-curve
The Secret World
Service level agreement
Sharing economy
Shockley, William
Short Message Service
Simple Mail Transfer Protocol
Six degrees of separation
Skype
Smartphones
Snapchat
Social media
Social network service
SONY
Spotify
Stakeholder Relationship Model
Standards
Star Wars The Old Republic
Starcraft 2
Stream Control Transmission Protocol
Subscription
Substitute
Switching costs
T
Telecommunications
industry
service
Telenor
Telia
3G
Tidal
Toshiba T1100
Tragedy of the commons
Transaction cost
Transistor
Transmission Control Protocol (TCP)
Tulip mania
Twitter
U
Uber
Uniform Resource Locator (URL)
United Nations (UN)
User Datagram Protocol (UDP)
V
Value
chain
generation
network
proposition
 shop
Video cassette recorder (VCR)
Video home system
Virtual network operator
mobile
Virtualization
Voice-over-IP
Vortex
W
WeChat
WhatsApp
WiFi
Wikimedia Foundation
Wikipedia
World of Warcraft
World Wide Web (WWW)
Y
YouTube
Z
Zero ARPU
Zero marginal cost
Zipf's law